@mariozechner/pi-agent-core
0.73.1 · 7 May 2026MIT523.2k / weekRepositoryHomepagenpmAtom feed
235 stable releases, newest first, by release line. Notes are quoted as the maintainers published them.
- 0.73.x · latest line
0.73.1
7 May 2026Self-update support for the npm scope migration:
pi update --selfnow supports the upcoming package rename from@mariozechner/pi-coding-agentto@earendil-works/pi-coding-agent. After the new package is published, existing global installs can update through the normal self-update flow; pi will uninstall the old global package and install the package name returned by the version check endpoint.Interactive OAuth login selection: OAuth providers can now present multiple login choices in
/login, enabling provider-specific interactive authentication flows. See Providers.JSONC-style
models.jsonparsing:models.jsonnow allows comments and trailing commas, making custom provider and model configuration easier to maintain. See Providers and Custom Providers.
New Features
Self-update support for the npm scope migration:
pi update --selfnow supports the upcoming package rename from@mariozechner/pi-coding-agentto@earendil-works/pi-coding-agent. After the new package is published, existing global installs can update through the normal self-update flow; pi will uninstall the old global package and install the package name returned by the version check endpoint.Interactive OAuth login selection: OAuth providers can now present multiple login choices in
/login, enabling provider-specific interactive authentication flows. See Providers.JSONC-style
models.jsonparsing:models.jsonnow allows comments and trailing commas, making custom provider and model configuration easier to maintain. See Providers and Custom Providers.
Added
Added interactive login selection support so OAuth providers can present multiple login choices
Changed
Changed
pi update --selfto honor the active package name returned by the Pi version check endpoint, defaulting to the current package when omitted and uninstalling the old global package before installing a renamed package.Changed extension loading to use upstream
jiti2.7 instead of the@mariozechner/jitiforkChanged
models.jsonparsing to allow comments and trailing commas
Fixed
Fixed
pi -ptreating prompts that start with YAML frontmatter as extension flags instead of user messages (#4163).Fixed pending tool results not updating in the live TUI after toggling thinking block visibility while the tool is running (#4167).
Fixed
/copyreporting success on Linux without writing the clipboard on Wayland-only compositors (Hyprland, Niri, ...) by skipping the X11-only native addon on Linux and routing throughwl-copy/xclip/xselinstead (#4177).Fixed HTML session exports to strip skill wrapper XML from rendered user messages
Fixed OpenAI-compatible chat completion streams that interleave content and tool-call deltas in the same choice.
Fixed OpenAI Codex OAuth refresh failures writing directly to stderr while the TUI is active (#4141).
Fixed OpenAI Codex Responses requests to send a non-empty system prompt (#4184).
Fixed Kimi For Coding model resolution for the Kimi K2 P6 alias (#4218).
Fixed Kitty inline image redraws to stay within TUI-owned terminal regions and avoid writing below the active viewport.
Fixed Kitty inline image rendering by letting the terminal allocate image ids and bounding parsed image ids to valid values.
Fixed inline image capability detection to disable inline images in cmux terminals.
0.73.0
4 May 2026- Breaking
Switched the built-in
xiaomiprovider from Token Plan AMS to Xiaomi's API billing endpoint, and renamed its/logindisplay from "Xiaomi MiMo Token Plan" to "Xiaomi MiMo".XIAOMI_API_KEYnow refers to the API billing key from platform.xiaomimimo.com. Users on Token Plan should switch to the appropriatexiaomi-token-plan-*provider and set the corresponding env var Xiaomi MiMo API billing and regional Token Plan providers -
xiaominow uses API billing, with separatexiaomi-token-plan-{cn,ams,sgp}providers. See docs/providers.md#api-keys and README.md#providers--models.Incremental bash output streaming - Bash tool output now appears while commands run instead of only after completion.
New Features
Xiaomi MiMo API billing and regional Token Plan providers -
xiaominow uses API billing, with separatexiaomi-token-plan-{cn,ams,sgp}providers. See docs/providers.md#api-keys and README.md#providers--models.Incremental bash output streaming - Bash tool output now appears while commands run instead of only after completion.
Compact read rendering - Interactive
readoutput for Pi docs, context files, and skills is collapsed by default and shows selected line ranges.
Breaking Changes
- Breaking
Switched the built-in
xiaomiprovider from Token Plan AMS to Xiaomi's API billing endpoint, and renamed its/logindisplay from "Xiaomi MiMo Token Plan" to "Xiaomi MiMo".XIAOMI_API_KEYnow refers to the API billing key from platform.xiaomimimo.com. Users on Token Plan should switch to the appropriatexiaomi-token-plan-*provider and set the corresponding env var
Added
Added three Xiaomi MiMo Token Plan regional providers visible in
/login:xiaomi-token-plan-cn(XIAOMI_TOKEN_PLAN_CN_API_KEY),xiaomi-token-plan-ams(XIAOMI_TOKEN_PLAN_AMS_API_KEY),xiaomi-token-plan-sgp(XIAOMI_TOKEN_PLAN_SGP_API_KEY). Each defaults tomimo-v2.5-pro
Changed
Changed
readtool rendering to collapse Pi documentation, AGENTS/CLAUDE context files, andSKILL.mdcontents by default in interactive output.
Fixed
Fixed generated OpenAI-compatible model metadata for Qwen 3.5/3.6 and MiniMax M2.7, so those models work through the built-in provider catalog
Fixed Bedrock Claude Opus 4.7
xhighthinking requests by preserving the provider's native effort value.Fixed OpenAI Codex WebSocket transport to fall back to SSE when setup fails before streaming starts, and surface transport diagnostics in the assistant message (#4133).
Fixed OpenAI Codex WebSocket transport keeping
--printand JSON mode processes alive after the response by closing cached WebSocket sessions during session shutdown (#4103).Fixed compact
readtool calls to render directly and include selected line ranges in interactive output.Fixed interactive sessions to exit when terminal input is lost instead of continuing in a broken state.
Fixed bash tool output to stream incrementally while commands run instead of waiting for command completion (#4145).
Fixed selector and autocomplete fuzzy ranking to prioritize exact matches.
- 0.72.x
0.72.1
2 May 2026Reading the note…
0.72.0
1 May 2026Reading the note…
- 0.71.x
0.71.1
1 May 2026Reading the note…
0.71.0
30 Apr 2026Reading the note…
- 0.70.x
0.70.6
28 Apr 2026Cloudflare Workers AI provider support with
CLOUDFLARE_API_KEY/CLOUDFLARE_ACCOUNT_IDsetup. See docs/providers.md#api-keys.Pi update checks now use
pi.devand identify Pi with api/<version>user agent. See docs/packages.md.Added Cloudflare Workers AI as a built-in provider with
CLOUDFLARE_API_KEY/CLOUDFLARE_ACCOUNT_IDsetup, default model resolution,/loginsupport, and provider documentation
New Features
Cloudflare Workers AI provider support with
CLOUDFLARE_API_KEY/CLOUDFLARE_ACCOUNT_IDsetup. See docs/providers.md#api-keys.Pi update checks now use
pi.devand identify Pi with api/<version>user agent. See docs/packages.md.
Added
Added Cloudflare Workers AI as a built-in provider with
CLOUDFLARE_API_KEY/CLOUDFLARE_ACCOUNT_IDsetup, default model resolution,/loginsupport, and provider documentation
Changed
Changed Pi version checks to identify Pi with a
pi/<version>user agent
Fixed
Fixed config selector scroll indicators to show item counts instead of line counts
Fixed exported HTML to escape embedded image data and session metadata, preventing crafted session content from injecting markup
Fixed Bun-based package manager startup by locating global
node_modulesrelative to Bun's install layoutFixed Bedrock inference profile capability checks by normalizing profile ARNs to the underlying model name.
Fixed file discovery to fall back to
fdfindwhenfdis unavailable.Fixed
pi updateto skip self-update reinstalls when the installed version is already current (#3853).Fixed Cloudflare Workers AI attribution headers to honor the install telemetry setting.
Fixed
pi update --selfdetection and execution for Windows package-manager shim installs, including symlinked global package roots, and print the manual fallback command when self-update fails (#3857).
0.70.5
27 Apr 2026Fixed HTML export preserving ANSI-renderer trailing padding as extra blank wrapped lines.
0.70.4
27 Apr 2026Fixed packaged
pistartup failing because the session selector imported a source-only utility path.
0.70.3
27 Apr 2026pi updatecan now update pi itself in addition to installed pi packages. See docs/packages.md.Azure Cognitive Services endpoint support for Azure OpenAI Responses deployments. See docs/providers.md#api-keys.
Suppressible Anthropic extra-usage billing warning via
warnings.anthropicExtraUsagein/settings. See docs/settings.md.
New Features
pi updatecan now update pi itself in addition to installed pi packages. See docs/packages.md.Azure Cognitive Services endpoint support for Azure OpenAI Responses deployments. See docs/providers.md#api-keys.
Suppressible Anthropic extra-usage billing warning via
warnings.anthropicExtraUsagein/settings. See docs/settings.md.Extension-controlled working row visibility via
ctx.ui.setWorkingVisible(), allowing extensions to hide the built-in loader row and render custom working state. See docs/extensions.md and examples/extensions/border-status-editor.ts.
Added
Added
pi updatesupport for updating pi itself in addition to installed pi packagesAdded Azure Cognitive Services endpoint support for Azure OpenAI Responses base URLs
Added
warnings.anthropicExtraUsageand a/settingswarnings submenu to suppress the Anthropic extra usage billing warningAdded
ctx.ui.setWorkingVisible()so extensions can hide the built-in interactive working loader row without reserving layout space, plus a border-status editor example that moves working state into a custom editor border
Fixed
Fixed duplicate printable characters from Kitty keyboard protocol CSI-u plus raw character input on layouts such as Italian (#3780).
Fixed API-key environment discovery and Bun startup to fall back to
/proc/self/environwhen Bun's sandbox leavesprocess.envemptyFixed Bun sandboxed package-manager commands when
process.envis emptyFixed symlinked packages, resources, skills, and sessions being duplicated in selectors and loaders
Fixed Bedrock prompt-caching and adaptive-thinking capability checks for inference profile ARNs
Fixed OpenAI Codex Responses default verbosity to
lowwhen no verbosity is specified.Stopped sending empty
toolsarrays to providers that reject them when tools are disabledFixed Anthropic SSE parsing to ignore unknown proxy events such as OpenAI-style
doneterminators (#3708).Fixed provider registration with override-only
models.jsonentries to preserve built-in model lists (#3651).Fixed
/loginto show auth supplied bymodels.jsonprovider definitions.Fixed HTML export whitespace around extension-rendered tool output and expandable output hints.
Fixed bash executor temp output streams leaking file descriptors when output was truncated by line count
Fixed extension
pi.setSessionName()updates to refresh the interactive terminal title immediatelyFixed
/treecancellation viasession_before_treeleaving the session stuck in compaction stateFixed Escape interrupt handling when extensions hide the built-in working loader row
Fixed coding-agent test expectations for current default models and missing-auth guidance.
Fixed long local-LLM SSE streams aborting at 5 minutes with
UND_ERR_BODY_TIMEOUTby disabling undicibodyTimeout/headersTimeouton the global dispatcher; provider SDKs continue to enforce their own deadlines viaretry.provider.timeoutMs
0.70.2
24 Apr 2026Fixed provider retry/timeout forwarding to omit undefined provider request controls, avoiding downstream SDK validation errors such as
timeout must be an integerwhenretry.provider.timeoutMsis not configured
0.70.1
24 Apr 2026DeepSeek provider support with V4 Flash/Pro models and
DEEPSEEK_API_KEYauthentication. See README.md#providers--models and docs/providers.md#api-keys.Provider request timeout/retry controls via
retry.provider.{timeoutMs,maxRetries,maxRetryDelayMs}, useful for long-running local inference and provider SDK retry behavior. See docs/settings.md#retry.Added DeepSeek to built-in provider setup, default model resolution, and provider documentation.
New Features
DeepSeek provider support with V4 Flash/Pro models and
DEEPSEEK_API_KEYauthentication. See README.md#providers--models and docs/providers.md#api-keys.Provider request timeout/retry controls via
retry.provider.{timeoutMs,maxRetries,maxRetryDelayMs}, useful for long-running local inference and provider SDK retry behavior. See docs/settings.md#retry.
Added
Added DeepSeek to built-in provider setup, default model resolution, and provider documentation.
Fixed
Fixed
/copyto avoid unbounded OSC 52 writes and clipboard races that could break terminal rendering or panic the native clipboard addonFixed extension flag docs to show
pi.getFlag()using registered flag names without the CLI--prefixFixed provider retry/timeout settings wiring by adding
retry.provider.{timeoutMs,maxRetries,maxRetryDelayMs}, migrating legacyretry.maxDelayMs, and forwarding provider controls intostreamSimplerequest optionsFixed Windows git package installs to bypass
cmd.exefor native git commands, so install paths containing spaces no longer breakpi install git:...withfatal: Too many argumentsFixed DeepSeek V4 session replay 400 errors by sending DeepSeek-compatible thinking controls and replayed assistant
reasoning_contentfieldsFixed GPT-5.5 generated context window metadata to use the observed 272k limit.
Fixed CSI-u Ctrl+letter decoding inside bracketed paste, so pasted modified-key escape sequences no longer become literal editor text
0.70.0
23 Apr 2026- Breaking
Disabled OSC 9;4 terminal progress indicators by default. Set
terminal.showTerminalProgresstotruein/settingsto re-enable Searchable auth provider login flow: the
/loginprovider selector now supports fuzzy search/filtering, making it faster to find providers when many are configured. See docs/providers.md.GPT-5.5 Codex support:
openai-codex/gpt-5.5is available as a model option, includingxhighreasoning support and corrected priority-tier pricing.
New Features
Searchable auth provider login flow: the
/loginprovider selector now supports fuzzy search/filtering, making it faster to find providers when many are configured. See docs/providers.md.GPT-5.5 Codex support:
openai-codex/gpt-5.5is available as a model option, includingxhighreasoning support and corrected priority-tier pricing.Terminal progress indicators are now opt-in: OSC 9;4 progress reporting during streaming/compaction is off by default and can be toggled via
terminal.showTerminalProgressin/settings--no-builtin-tools/createAgentSession({ noTools: "builtin" })now correctly disables only built-in tools while keeping extension tools active. See docs/extensions.md and README.md
Breaking Changes
- Breaking
Disabled OSC 9;4 terminal progress indicators by default. Set
terminal.showTerminalProgresstotruein/settingsto re-enable
Added
Added searchable auth provider login flow with fuzzy filtering in the provider selector
Added GPT-5.5 Codex model
Added auth source labels in
/loginso provider entries can show when auth comes from--api-key, an environment variable, or custom provider fallback without exposing secrets.
Changed
Updated default model selection across providers to current recommended models.
Improved stale extension context errors after session replacement or reload to tell extension authors to avoid captured
pi/commandctxand usewithSessionfor post-replacement work.
Fixed
Fixed
/modelselector cancellation to request render instead of incorrectly triggering login selector.Changed login, OAuth, and extension selectors for more consistent styling.
Added Amazon Bedrock setup guidance to
/loginand updated/modelcopy to refer to configured providers instead of only API keys.Improved no-model and missing-auth warnings to point users to
/loginfor OAuth or API key setup.Fixed
/quitshutdown ordering to stop the TUI before extension UI teardown can repaint, preserving the final rendered frame while still emittingsession_shutdownbefore process exit.Fixed
SettingsManager.inMemory()initial settings being lost after reloads triggered by SDK resource loadingFixed
models.jsonprovider compatibility to acceptcompat.supportsLongCacheRetention, allowing proxies to opt out of long-retention cache fields when needed while long retention is enabled by default when requestedFixed
--thinking xhighforopenai-codexgpt-5.5so it is no longer downgraded tohigh.Fixed git package installs with custom
npmCommandvalues such aspnpmby avoiding npm-specific production flags in that compatibility pathFixed first user messages rendering without spacing after existing notices such as compaction summaries or status messages
Fixed the handoff extension example to use the replacement-session context after creating a new session, avoiding stale
ctxerrors when it installs the generated promptFixed session replacement and
/quitteardown ordering to run host-owned extension UI cleanup synchronously aftersession_shutdownhandlers complete but before invalidating the old extension context, preventing stale extension UI from rendering against a disposed sessionFixed crash on
/quitwhen an extension registers a custom footer whoserender()accessesctx, by tearing down extension-provided UI before invalidating the extension runner during shutdownFixed auto-retry to treat Bedrock/Smithy HTTP/2 transport failures like
http2 request did not get a responseas transient errors, so the agent retries automatically instead of waiting for a manual nudgeFixed the CLI/SDK tool-selection split so
--no-builtin-toolsandcreateAgentSession({ noTools: "builtin" })disable only built-in default tools while keeping extension/custom tools enabled, instead of falling through to the same "disable everything" path as--no-toolsFixed remaining hardcoded
pi/.pibranding to route throughAPP_NAMEandCONFIG_DIR_NAMEextension points, so SDK rebrands get consistent naming in/quitdescription,process.title, and the project-local extensions directoryFixed
pi-coding-agentshippinguuid@11, which triggerednpm auditmoderate vulnerability reports for downstream installs; the package now depends onuuid@14Fixed
openai-completionsstreamed tool-call assembly to coalesce deltas by stable tool index when OpenAI-compatible gateways mutate tool call IDs mid-stream, preventing malformed Kimi K2.6/OpenCode tool streams from splitting one call into multiple bogus tool callsFixed
ctx.ui.setWorkingMessage()to persist across loader recreation, matching the behavior ofctx.ui.setWorkingIndicator()Fixed coding-agent
fs.watcherror handling for theme and git-footer watchers to retry after transient watcher failures such asEMFILE, avoiding startup crashes in large reposFixed built-in
kimi-codingmodel generation to attach the expectedUser-Agentheader so direct Kimi Coding requests use the provider's expected client identityFixed extension shortcut conflict diagnostics to display at startup instead of only on reload, so extension authors discover reserved keybinding conflicts immediately rather than discovering them later through user feedback
Fixed
models.jsonAnthropic-compatible provider configuration to acceptcompat.supportsEagerToolInputStreaming, allowing proxies that reject per-tooleager_input_streamingto use the legacy fine-grained tool streaming beta header insteadFixed startup banner extension labels to strip trailing
index.js/index.tssuffixesFixed OSC 9;4 terminal progress updates to stay alive in terminals such as Ghostty during long-running agent work
Fixed OpenAI-compatible completion usage parsing to avoid double-counting reasoning tokens already included in
completion_tokensFixed
openai-responsescompatibility for strict OpenAI-compatible proxies by allowingmodels.jsonto disable the underscore-containingsession_idheader withcompat.sendSessionIdHeader: falseFixed GPT-5.5 Codex capability handling to clamp unsupported minimal reasoning to
lowand apply the model's 2.5x priority service-tier pricing multiplier
- 0.69.x
0.69.0
22 Apr 2026- Breaking
Migrated first-party coding-agent code, SDK/examples/docs, and package metadata from
@sinclair/typebox0.34.x totypebox1.x. New extensions, SDK integrations, and pi packages should depend on and import fromtypebox. Legacy extension loading still aliases the root@sinclair/typeboxpackage, but@sinclair/typebox/compileris no longer shimmed. This migration also picks up the new@mariozechner/pi-aiTypeBox-native validator path, so tool argument validation now works in eval-restricted runtimes such as Cloudflare Workers instead of being skipped - Breaking
Session-replacement commands now invalidate captured pre-replacement session-bound extension objects after
ctx.newSession(),ctx.fork(), andctx.switchSession(). Oldpiand commandctxreferences now throw instead of silently targeting the replaced session. Migration: if code needs to keep working in the replacement session after one of those calls, passwithSessionto that same method and do the post-switch work there. In practice, move post-switchpi.sendUserMessage(),pi.sendMessage(), and command-ctx/session-manager access intowithSession, and use only theReplacedSessionContextpassed to that callback for session-bound operations. Footguns:withSessionruns after the old extension instance has already receivedsession_shutdown, old cleanup may already have invalidated captured state, captured oldpi/ old commandctxare stale, and previously extracted raw objects such asconst sm = ctx.sessionManagerremain the caller's responsibility and must not be reused after the switch. TypeBox 1.x migration for extensions and SDK integrations, including TypeBox-native tool argument validation that now works in eval-restricted runtimes such as Cloudflare Workers. See docs/extensions.md and docs/sdk.md.
New Features
TypeBox 1.x migration for extensions and SDK integrations, including TypeBox-native tool argument validation that now works in eval-restricted runtimes such as Cloudflare Workers. See docs/extensions.md and docs/sdk.md.
Stacked extension autocomplete providers via
ctx.ui.addAutocompleteProvider(...), allowing extensions to layer custom completion logic on top of built-in slash and path completion. See docs/extensions.md#autocomplete-providers and examples/extensions/github-issue-autocomplete.ts.Terminating tool results via
terminate: true, allowing custom tools to end on a final tool call without paying for an automatic follow-up LLM turn. See docs/extensions.md and examples/extensions/structured-output.ts.OSC 9;4 terminal progress indicators during agent streaming and compaction for supporting terminals.
Breaking Changes
- Breaking
Migrated first-party coding-agent code, SDK/examples/docs, and package metadata from
@sinclair/typebox0.34.x totypebox1.x. New extensions, SDK integrations, and pi packages should depend on and import fromtypebox. Legacy extension loading still aliases the root@sinclair/typeboxpackage, but@sinclair/typebox/compileris no longer shimmed. This migration also picks up the new@mariozechner/pi-aiTypeBox-native validator path, so tool argument validation now works in eval-restricted runtimes such as Cloudflare Workers instead of being skipped - Breaking
Session-replacement commands now invalidate captured pre-replacement session-bound extension objects after
ctx.newSession(),ctx.fork(), andctx.switchSession(). Oldpiand commandctxreferences now throw instead of silently targeting the replaced session. Migration: if code needs to keep working in the replacement session after one of those calls, passwithSessionto that same method and do the post-switch work there. In practice, move post-switchpi.sendUserMessage(),pi.sendMessage(), and command-ctx/session-manager access intowithSession, and use only theReplacedSessionContextpassed to that callback for session-bound operations. Footguns:withSessionruns after the old extension instance has already receivedsession_shutdown, old cleanup may already have invalidated captured state, captured oldpi/ old commandctxare stale, and previously extracted raw objects such asconst sm = ctx.sessionManagerremain the caller's responsibility and must not be reused after the switch.
Added
Added support for terminating tool results via
terminate: true, allowing custom tools to end the current tool batch without an automatic follow-up LLM call, plus astructured-output.tsextension example and extension docs showing the patternAdded OSC 9;4 terminal progress indicators during agent streaming and compaction, so terminals like iTerm2, WezTerm, Windows Terminal, and Kitty show activity in their tab bar
Added
ctx.ui.addAutocompleteProvider(...)for stacking extension autocomplete providers on top of the built-in slash/path provider, plus agithub-issue-autocomplete.tsexample and extension docs
Fixed
Fixed exported session HTML to sanitize markdown link URLs before rendering them into anchor tags, blocking
javascript:-style payloads while preserving safe links in shared/exported sessionsFixed
ctx.getSystemPrompt()insidebefore_agent_startto reflect chained system-prompt changes made by earlierbefore_agent_starthandlers, and clarified the extension docs around provider-payload rewrites and whatctx.getSystemPrompt()does and does not reportFixed built-in
google-gemini-climodel lists and selector entries to includegemini-3.1-flash-lite-preview, so Cloud Code Assist users no longer need manual--modelfallback selection to use itFixed extension session-replacement flows so
ctx.newSession(),ctx.fork(),ctx.switchSession(), and imported-session replacements fully rebind before post-switch work runs, addedwithSessionreplacement callbacks with freshReplacedSessionContexthelpers, and make stale pre-replacementpi/ctxsession-bound accesses throw instead of silently targeting the wrong sessionFixed
models.jsonbuilt-in provider overrides to acceptheaderswithout requiringbaseUrl, so request-header-only overrides now load and apply correctly
- 0.68.x
0.68.1
22 Apr 2026Fireworks provider support with built-in models and
FIREWORKS_API_KEYauth. See README.md#providers--models and docs/providers.md.Configurable inline tool image width via
terminal.imageWidthCellsin/settings. See docs/settings.md#terminal--images.Added built-in Fireworks provider support, including
FIREWORKS_API_KEYsetup/docs and the default Fireworks modelaccounts/fireworks/models/kimi-k2p6
New Features
Fireworks provider support with built-in models and
FIREWORKS_API_KEYauth. See README.md#providers--models and docs/providers.md.Configurable inline tool image width via
terminal.imageWidthCellsin/settings. See docs/settings.md#terminal--images.
Added
Added built-in Fireworks provider support, including
FIREWORKS_API_KEYsetup/docs and the default Fireworks modelaccounts/fireworks/models/kimi-k2p6
Fixed
Fixed interactive inline tool images to honor configurable
terminal.imageWidthCellsvia/settings, so tool-output images are no longer hard-capped to 60 terminal cellsFixed
sessionDirinsettings.jsonto expand~, so portable session-directory settings no longer require a shell wrapperFixed parallel tool-call rows to leave the pending state as soon as each tool is finalized, while still appending persisted tool results in assistant source order
Fixed exported session markdown to render Markdown while showing HTML-like message content such as
<file name="...">...</file>verbatim, so shared sessions match the TUI instead of letting the browser interpret message textFixed exported session HTML to render
grepandfindoutput through their existing TUI renderers andlsoutput through a native template renderer, avoiding missing formatting and spacing artifacts in shared sessionsFixed
@autocomplete fuzzy search to follow symlinked directories and include symlinked paths in resultsFixed proxied agent streams to preserve the proxy-safe serializable subset of stream options, including session, transport, retry-delay, metadata, header, cache-retention, and thinking-budget settings
Hardened Anthropic streaming against malformed tool-call JSON by owning SSE parsing with defensive JSON repair, replacing the deprecated
fine-grained-tool-streamingbeta header with per-tooleager_input_streaming, and updating stale test model referencesFixed Bedrock runtime endpoint resolution to stop pinning built-in regional endpoints over
AWS_REGION/AWS_PROFILE, restoringus.*andeu.*inference profile support after v0.68.0 while preserving custom VPC/proxy endpoint overrides (#3481, #3485, #3486, #3487, #3488)
0.68.0
20 Apr 2026- Breaking
Changed SDK and CLI tool selection from cwd-bound built-in tool instances to tool-name allowlists.
createAgentSession({ tools })now expectsstringnames such as"read"and"bash"instead ofTool,--toolsnow allowlists built-in, extension, and custom tools by name, and--no-toolsnow disables all tools by default rather than only built-ins. Migrate SDK code fromtools: [readTool, bashTool]totools: ["read", "bash"](#2835, #3452) - Breaking
Removed prebuilt cwd-bound tool and tool-definition exports from
@mariozechner/pi-coding-agent, includingreadTool,bashTool,editTool,writeTool,grepTool,findTool,lsTool,readOnlyTools,codingTools, and the corresponding*ToolDefinitionvalues. Use the explicit factory exports instead, for examplecreateReadTool(cwd),createBashTool(cwd),createCodingTools(cwd), andcreateReadToolDefinition(cwd) - Breaking
Removed ambient
process.cwd()/ default agent-dir fallback behavior from public resource helpers.DefaultResourceLoader,loadProjectContextFiles(), andloadSkills()now require explicit cwd/agent-dir style inputs, and exported system-prompt option types now require an explicitcwd. Pass the session or project cwd explicitly instead of relying on process-global defaults
New Features
Configurable streaming working indicator for extensions via
ctx.ui.setWorkingIndicator(), including animated, static, and hidden indicators. See docs/tui.md#working-indicator, docs/extensions.md, and examples/extensions/working-indicator.ts.before_agent_startnow exposessystemPromptOptions(BuildSystemPromptOptions) so extensions can inspect the structured system-prompt inputs without re-discovering resources. See docs/extensions.md#before_agent_start and examples/extensions/prompt-customizer.ts.Configurable keybindings for scoped model selector actions and session-tree filter actions. See docs/keybindings.md.
/cloneduplicates the current active branch into a new session, while extensions can choose whether to forkbeforeoratan entry viactx.fork(..., { position }). See README.md, docs/extensions.md, and docs/session.md.
Breaking Changes
- Breaking
Changed SDK and CLI tool selection from cwd-bound built-in tool instances to tool-name allowlists.
createAgentSession({ tools })now expectsstringnames such as"read"and"bash"instead ofTool,--toolsnow allowlists built-in, extension, and custom tools by name, and--no-toolsnow disables all tools by default rather than only built-ins. Migrate SDK code fromtools: [readTool, bashTool]totools: ["read", "bash"](#2835, #3452) - Breaking
Removed prebuilt cwd-bound tool and tool-definition exports from
@mariozechner/pi-coding-agent, includingreadTool,bashTool,editTool,writeTool,grepTool,findTool,lsTool,readOnlyTools,codingTools, and the corresponding*ToolDefinitionvalues. Use the explicit factory exports instead, for examplecreateReadTool(cwd),createBashTool(cwd),createCodingTools(cwd), andcreateReadToolDefinition(cwd) - Breaking
Removed ambient
process.cwd()/ default agent-dir fallback behavior from public resource helpers.DefaultResourceLoader,loadProjectContextFiles(), andloadSkills()now require explicit cwd/agent-dir style inputs, and exported system-prompt option types now require an explicitcwd. Pass the session or project cwd explicitly instead of relying on process-global defaults
Added
Added extension support for customizing the interactive streaming working indicator via
ctx.ui.setWorkingIndicator(), including custom animated frames, static indicators, hidden indicators, a newworking-indicator.tsexample extension, and updated extension/TUI/RPC docsAdded
systemPromptOptions(BuildSystemPromptOptions) tobefore_agent_startextension events, so extensions can inspect the structured inputs used to build the current system promptAdded
/cloneto duplicate the current active branch into a new session, while keeping/forkfocused on forking from a previous user messageAdded
ctx.fork()support forposition: "before" | "at"so extensions and integrations can branch before a user message or duplicate the current point in the conversation; the interactive clone/fork UX builds on that runtime supportAdded configurable keybinding ids for scoped model selector actions and tree filter actions, so those interactive shortcuts can be remapped in
keybindings.jsonAdded
PI_OAUTH_CALLBACK_HOSTsupport for built-in OAuth login flows, allowing local callback servers used bypi authto bind to a custom interface instead of hardcoded127.0.0.1Added
reasonandtargetSessionFilemetadata tosession_shutdownextension events, so extensions can distinguish quit, reload, new-session, resume, and fork teardown paths
Changed
Changed
pi updateto batch npm package updates per scope and run git package updates with bounded parallelism, reducing multi-package update time while preserving skip behavior for pinned and already-current packagesChanged Bedrock session requests to omit
maxTokenswhen model token limits are unknown and to omittemperaturewhen unset, letting Bedrock use provider defaults and avoid unnecessary TPM quota reservation
Fixed
Fixed
AgentSessionsystem-prompt option initialization to avoid constructing an invalid emptyBuildSystemPromptOptions, sonpm run checkpasses aftercwdbecame mandatory.Fixed shell-path resolution to stop consulting ambient
process.cwd()state during bash execution, so session/project-specificshellPathsettings now follow the active coding-agent session cwd instead of the launcher cwdFixed
ctx.ui.setWorkingIndicator()custom frames to render verbatim instead of forcing the theme accent color, so extensions now own working-indicator coloring when they customize itFixed
pi updatereinstalling npm packages that are already at the latest published version by checking the installed package version before runningnpm install <pkg>@latestFixed
@autocomplete plain queries to stop matching against the full cwd/base path, so path fragments in worktree names no longer crowd out intended results such as@planFixed built-in tool wrapping to use the same extension-runner context path as extension tools, so built-in tools receive execution context and
readcan warn when the current model does not support imagesFixed
openai-completionsassistant replay to preservecompat.requiresThinkingAsTexttext-part serialization, avoiding same-model follow-up crashes when previous assistant messages mix thinking and textFixed direct OpenAI Chat Completions sessions to map
sessionIdandcacheRetentionto prompt caching fields, sendingprompt_cache_keywhen caching is enabled andprompt_cache_retention: "24h"for directapi.openai.comrequests with long retentionFixed OpenAI-compatible Chat Completions sessions to optionally send aligned
session_id,x-client-request-id, andx-session-affinityheaders fromsessionIdviacompat.sendSessionAffinityHeaders, improving cache-affinity routing for backends such as FireworksFixed threaded
/resumesession relationships and current-session detection to canonicalize symlinked session paths during selector comparisons, so shared session directories no longer break parent-child matching or active-session delete protectionFixed
/session, Sessions docs, and CLI help to consistently document that session reuse supports both file paths and session IDs, and that/sessionshows the current session IDFixed Windows pnpm global install detection to recognize
\\.pnpm\\store paths, so update notices now suggestpnpm install -g @mariozechner/pi-coding-agentinstead of falling back to npmFixed missing
@sinclair/typeboxruntime dependency in@mariozechner/pi-coding-agent, so strict pnpm installs no longer fail withERR_MODULE_NOT_FOUNDwhen startingpiFixed xterm uppercase typing in the interactive editor by decoding printable
modifyOtherKeysinput and normalizing shifted letter matching, soShift+letterno longer disappears inpiFixed
/compactto reuse the session thinking level for compaction summaries instead of forcinghigh, avoiding invalid reasoning-effort errors ongithub-copilot/claude-opus-4.7sessions configured formediumthinkingFixed shared/exported plain-text tool output to preserve indentation instead of collapsing leading whitespace in the web share page
Fixed exported share pages to use browser-safe
TandOshortcuts with clickable header toggles for thinking and tool visibility instead of browser-reservedCtrl+T/Ctrl+ObindingsFixed skill resolution to dedupe symlinked aliases by canonical path, so
pi configno longer shows duplicate skill entries when~/.pi/agent/skillspoints to~/.agents/skillsFixed OpenRouter request attribution to include Pi app headers (
HTTP-Referer: https://pi.dev,X-OpenRouter-Title: pi,X-OpenRouter-Categories: cli-agent) when sessions are created through the coding-agent SDK and install telemetry is enabledFixed custom-model
compatschema/docs to supportcacheControlFormat: "anthropic"for OpenAI-compatible providers that expose Anthropic-style prompt caching viacache_controlmarkersFixed Cloud Code Assist tool schemas to strip JSON Schema meta-declaration keys before provider translation, avoiding validation failures for tool-enabled sessions that use
$schema,$defs, and related metadataFixed direct Bedrock sessions to honor
model.baseUrlas the runtime client endpoint, restoring support for custom Bedrock VPC or proxy routesFixed the
edittool to coerce stringifiededitsJSON before validation, so models that send the array payload as a JSON string no longer fall back to ad-hoc shell editsFixed package manifest positive glob entries to expand before loading packaged resources, restoring manifest patterns such as
skills/**/*.md
- 0.67.x
0.67.68
17 Apr 20260.67.67
17 Apr 2026Bedrock sessions can now authenticate with
AWS_BEARER_TOKEN_BEDROCK, enabling Converse API access without local SigV4 credentials. See docs/providers.md#amazon-bedrock.Added Bedrock bearer-token authentication support via
AWS_BEARER_TOKEN_BEDROCK, enabling coding-agent sessions to use Bedrock Converse without local SigV4 credentialsFixed
/scoped-modelsAlt+Up/Down to stay a no-op in the implicitall enabledstate instead of materializing a full explicit enabled-model list and marking the selector dirty
New Features
Bedrock sessions can now authenticate with
AWS_BEARER_TOKEN_BEDROCK, enabling Converse API access without local SigV4 credentials. See docs/providers.md#amazon-bedrock.
Added
Added Bedrock bearer-token authentication support via
AWS_BEARER_TOKEN_BEDROCK, enabling coding-agent sessions to use Bedrock Converse without local SigV4 credentials
Fixed
Fixed
/scoped-modelsAlt+Up/Down to stay a no-op in the implicitall enabledstate instead of materializing a full explicit enabled-model list and marking the selector dirtyFixed Mistral Small 4 default thinking requests to use the model's supported reasoning control, avoiding
400errors when starting sessions onmistral-small-2603andmistral-small-latestFixed Qwen chat-template thinking replay to preserve prior thinking across turns, so affected OpenAI-compatible models keep multi-turn tool-call arguments instead of degrading to empty
{}payloadsFixed exported HTML transcripts so text selection no longer triggers click-based expand/collapse toggles
Fixed flaky git package update notifications by waiting for captured git command stdio to fully drain before comparing local and remote commit SHAs
Fixed system prompt dates to use a stable
YYYY-MM-DDformat instead of locale-dependent output, keeping prompts deterministic across runtimes and localesFixed auto-retry transient error detection to treat
Network connection lost.as retryable, so dropped provider connections retry instead of terminating the agentFixed compact interactive extension startup summaries to disambiguate package extensions and repeated local
index.tsentries by using package-aware labels and the minimal parent path needed to make local entries uniqueFixed git package dependency installation to use production installs (
npm install --omit=dev) during both install and update flows, so extension runtime dependencies must come fromdependenciesand notdevDependenciesFixed
tool_result/afterToolCallextension handling for error results by forwardingdetailsandisErroroverrides throughAgentSessioninstead of dropping them whenisErrorwas already trueFixed missing root exports for
RpcClientand RPC protocol types from@mariozechner/pi-coding-agent, so ESM consumers can import them from the main package entrypointFixed OpenAI Codex service-tier cost accounting to trust the explicitly requested tier when the API echoes the default tier in responses, keeping session cost displays aligned with the selected tier
Fixed parallel tool-call finalization to convert
afterToolCallhook throws into error tool results instead of aborting the remaining tool batchFixed Bun binary asset path resolution to honor
PI_PACKAGE_DIRfor built-in themes, HTML export templates, and interactive bundled assetsFixed user-message turn spacing in interactive mode by restoring an inter-message spacer before user turns (except the first user message), preventing assistant and user blocks from rendering flush together.
Fixed interactive
/importhandling to support quoted JSONL paths with spaces, route missing JSONL files through the non-fatalSessionImportFileNotFoundErrorpath, and document theimportFromJsonl()exceptions (SessionImportFileNotFoundError,MissingSessionCwdError).
0.67.6
16 Apr 2026Prompt templates support an
argument-hintfrontmatter field that renders before the description in the/autocomplete dropdown, using<angle>for required and[square]for optional arguments. See docs/prompt-templates.md#argument-hints.New
after_provider_responseextension hook lets extensions inspect provider HTTP status codes and headers immediately after each response is received and before stream consumption begins. See docs/extensions.md.Compact interactive startup header with a comma-separated view of loaded AGENTS.md files, prompt templates, skills, and extensions. Press
Ctrl+Oto toggle the expanded listing.
New Features
Prompt templates support an
argument-hintfrontmatter field that renders before the description in the/autocomplete dropdown, using<angle>for required and[square]for optional arguments. See docs/prompt-templates.md#argument-hints.New
after_provider_responseextension hook lets extensions inspect provider HTTP status codes and headers immediately after each response is received and before stream consumption begins. See docs/extensions.md.Compact interactive startup header with a comma-separated view of loaded AGENTS.md files, prompt templates, skills, and extensions. Press
Ctrl+Oto toggle the expanded listing.Markdown links in assistant output now render as OSC 8 hyperlinks on terminals that advertise support; unknown terminals and tmux/screen default to plain text so URLs are never silently dropped.
Added
Added
argument-hintfrontmatter field for prompt templates, displayed before the description in the autocomplete dropdownAdded
after_provider_responseextension hook so extensions can inspect provider HTTP status codes and headers after each provider response is received and before stream consumption beginsAdded OSC 8 hyperlink rendering for markdown links when the terminal advertises support
Changed
Changed interactive startup header to a compact, comma-separated view of loaded AGENTS.md files, prompt templates, skills, and extensions, with
Ctrl+Oto toggle the expanded listingTightened hyperlink capability detection to default
hyperlinks: falsefor unknown terminals and force it off under tmux/screen (including nested sessions), preventing markdown link URLs from disappearing on terminals that silently swallow OSC 8 sequences
Fixed
Fixed
--verbosestartup output to begin with expanded startup help and loaded resource listings after the compact startup header changeFixed
findtool returning no results for path-based glob patterns such assrc/**/*.spec.tsorsome/parent/child/**by switching fd into full-path mode and normalizing the pattern when it contains a/Fixed
findtool applying nested.gitignorerules across sibling directories (e.g. rules froma/.gitignorehiding matching files underb/) by dropping the manual--ignore-filecollection and delegating to fd's hierarchical.gitignorehandling via--no-require-gitFixed OpenAI Responses prompt caching for non-
api.openai.combase URLs (OpenAI-compatible proxies such as litellm, theclawbay) by sending thesession_idandx-client-request-idcache-affinity headers unconditionally when asessionIdis provided, matching the official Codex CLI behaviorFixed the
presetexample extension to snapshot the active model, thinking level, and tool set on the first preset application and restore that state when cycling back to(none), instead of falling back to a hardcoded default tool list
0.67.5
16 Apr 2026Fixed Opus 4.7 adaptive thinking configuration across Anthropic and Bedrock providers by recognizing Opus 4.7 adaptive-thinking support and mapping
xhighreasoning to provider-supported effort valuesFixed Zellij
Shift+Enterregressions by reverting the Zellij-specific Kitty keyboard query bypass and restoring the previous keyboard negotiation behavior
0.67.4
16 Apr 2026--no-context-files(-nc) disables automaticAGENTS.md/CLAUDE.mddiscovery when you need a clean run without project context injection. See README.md#context-files.loadProjectContextFiles()is now exported as a standalone utility for extensions and SDK-style integrations that need to inspect the same context-file resolution order used by the CLI. See README.md#context-files.New
after_provider_responseextension hook lets extensions inspect provider HTTP status codes and headers immediately after response creation and before stream consumption. See docs/extensions.md.
New Features
--no-context-files(-nc) disables automaticAGENTS.md/CLAUDE.mddiscovery when you need a clean run without project context injection. See README.md#context-files.loadProjectContextFiles()is now exported as a standalone utility for extensions and SDK-style integrations that need to inspect the same context-file resolution order used by the CLI. See README.md#context-files.New
after_provider_responseextension hook lets extensions inspect provider HTTP status codes and headers immediately after response creation and before stream consumption. See docs/extensions.md.
Added
Added
--no-context-files(-nc) to disableAGENTS.mdandCLAUDE.mdcontext file discovery and loadingExported
loadProjectContextFiles()as a standalone utility so extensions can discover project context files without instantiating a fullDefaultResourceLoaderAdded
after_provider_responseextension hook so extensions can inspect provider HTTP status codes and headers after each provider response is received and before stream consumption begins
Changed
Added
claude-opus-4-7model for Anthropic.Changed Anthropic prompt caching to add a
cache_controlbreakpoint on the last tool definition, so tool schemas can be cached independently from transcript updates while preserving existing cache retention behavior
Fixed
Fixed markdown strikethrough parsing in interactive rendering and HTML export to require strict double-tilde delimiters (
~~text~~) with non-whitespace boundaries.Fixed shutdown handling to kill tracked detached
bashtool child processes on exit signals, preventing orphaned background processes.Fixed flaky
edit-tool-no-full-redrawTUI tests by waiting for asynchronous preview and preflight error rendering instead of relying on fixed render ticks.Fixed
kimi-codingdefault model selection to usekimi-for-codinginstead ofkimi-k2-thinkingFixed
ctrl+zon native Windows to avoid crashing interactive mode, disable the default suspend binding there, and show a status message when suspend is invoked manuallyFixed
findtool cancellation and responsiveness on broad searches by making.gitignorediscovery andfdexecution fully abort-aware and non-blockingFixed
grepbroad-search stalls whencontext=0by formatting match lines from ripgrep JSON output instead of doing synchronous per-match file reads
0.67.3
15 Apr 2026renderShell: "self"for custom and built-in tool renderers so tools can own their outer shell instead of the default boxed shell. Useful for stable large previews such as edit diffs. See docs/extensions.md#custom-rendering.Interactive auto-retry status now shows a live countdown during backoff instead of a static retry delay message.
Added
renderShell: "self"for custom and built-in tool renderers so tools can own their outer shell instead of using the default boxed shell. This is useful for stable large previews such as edit diffs
New Features
renderShell: "self"for custom and built-in tool renderers so tools can own their outer shell instead of the default boxed shell. Useful for stable large previews such as edit diffs. See docs/extensions.md#custom-rendering.Interactive auto-retry status now shows a live countdown during backoff instead of a static retry delay message.
Added
Added
renderShell: "self"for custom and built-in tool renderers so tools can own their outer shell instead of using the default boxed shell. This is useful for stable large previews such as edit diffs
Fixed
Fixed edit diff previews to stay visible during edit permission dialogs and session replay without reintroducing large-result redraw flicker
Fixed
/reloadto render a static reload status box instead of an animated spinner, avoiding redraw instability during interactive reloads.Fixed the
plan-modeexample extension to allowezain the read-only bash allowlist instead of the deprecatedexacommandFixed
google-vertexAPI key resolution to treatgcp-vertex-credentialsas an Application Default Credentials marker instead of a literal API key, so marker-based setups correctly fall back to ADCFixed RPC
promptto wait for prompt preflight success before emitting its single authoritative response, while still treating handled and queued prompts as successFixed Alt keybindings inside Zellij by skipping the Kitty keyboard protocol query there and enabling xterm
modifyOtherKeysmode 2 directlyFixed
/scoped-modelsreordering to propagate into the/modelscoped tab, preserving the user-defined scoped model order instead of re-sorting itFixed
session_shutdownto fire onSIGHUPandSIGTERMin interactive, print, and RPC modes so extensions can run shutdown cleanup on those signal-driven exitsFixed screenshot path parsing to handle lower case am/pm in macOS screenshot filenames
Fixed interactive auto-retry status updates to show a live countdown during backoff instead of a static retry delay message
0.67.2
14 Apr 2026Support for multiple
--append-system-promptflags, each value is appended to the system prompt separated by double newlines. See README.md#other-options.Support for passing inline extension factories to
main()for embedded integrations and custom entrypoints.Interactive keybinding support for Kitty
super-modified shortcuts such assuper+k,super+enter, andctrl+super+k. See docs/keybindings.md.
New Features
Support for multiple
--append-system-promptflags, each value is appended to the system prompt separated by double newlines. See README.md#other-options.Support for passing inline extension factories to
main()for embedded integrations and custom entrypoints.Interactive keybinding support for Kitty
super-modified shortcuts such assuper+k,super+enter, andctrl+super+k. See docs/keybindings.md.
Added
Added support for multiple
--append-system-promptflags, each value is appended to the system prompt separated by double newlinesAdded interactive keybinding support for Kitty
super-modified shortcuts such assuper+k,super+enter, andctrl+super+kAdded support for passing inline extension factories to
main()for embedded integrations and custom entrypoints
Fixed
Fixed direct OpenAI Responses and Codex SSE requests to align
prompt_cache_key,session_id, andx-client-request-idvalues with the same session-derived identifier, improving prompt cache affinity for append-only sessionsFixed streaming-only
partialJsonscratch buffers leaking into persisted OpenAI Responses tool calls, which could corrupt follow-up payloads on resumed conversations.Fixed Ctrl+Alt letter key matching in tmux by falling through from legacy ESC-prefixed handling to CSI-u and xterm
modifyOtherKeysparsing when the legacy form does not matchFixed the shipped
subagentexample to avoid leaking Bun virtual filesystem script paths into subagent promptsFixed bordered loaders to stop their animation timer when disposed, preventing stale loader updates after teardown.
0.67.1
13 Apr 2026Pi needs a reliable per-version usage signal to understand whether releases are being adopted and to help justify funding continued development.
Npm download counts are not a reliable proxy for actual Pi usage.
It only runs in interactive mode.
Telemetry
Interactive mode now sends a lightweight anonymous install/update telemetry ping to
https://pi.dev/install?version=x.y.zafter it writeslastChangelogVersioninsettings.json.Why this exists:
Pi needs a reliable per-version usage signal to understand whether releases are being adopted and to help justify funding continued development.
Npm download counts are not a reliable proxy for actual Pi usage.
How it works:
It only runs in interactive mode.
It does not run in RPC mode, print mode, JSON mode, or SDK mode.
On a fresh interactive install, Pi writes
lastChangelogVersion, then sends the ping.On later interactive startups, if the local changelog contains entries newer than the previously stored
lastChangelogVersion, Pi writes the newlastChangelogVersion, then sends the ping.The request is fire-and-forget. Startup does not wait for it, and any errors are ignored.
What data is collected:
Only the Pi version in the request path, for example
https://pi.dev/install?version=0.67.1.The server stores only aggregate per-version counters such as
{ "0.67.1": 3 }.It does not store IP addresses, client identifiers, prompts, paths, models, auth state, or any other per-user data. It literally only increments a counter for that version.
How to disable it:
/settings→ disableInstall telemetrysettings.json→ setenableInstallTelemetrytofalsePI_OFFLINE=1PI_TELEMETRY=0
New Features
Full
openRouterRoutingsupport inmodels.json, including fallbacks, parameter requirements, data collection, ZDR, ignore lists, quantizations, provider sorting, max price, and preferred throughput and latency constraints. See docs/models.md.PI_CODING_AGENT=trueenvironment variable set at startup so subprocesses can detect they are running inside the coding agent.
Added
Added full
openRouterRoutingfield support inmodels.json, including fallbacks, parameter requirements, data collection, ZDR, ignore lists, quantizations, provider sorting, max price, and preferred throughput and latency constraintsSet
PI_CODING_AGENT=trueenvironment variable at startup so sub-processes can detect they are running inside the coding agent
Fixed
Fixed interactive changelog rendering for the telemetry notes by moving the section under a
### Telemetryheading, so startup shows the full release notes instead of only the version header.Updated
antigravity-image-gen.tsexample extension to use User-Agent version1.21.9Bumped default Antigravity User-Agent version to
1.21.9Fixed Gemma 4 thinking level mapping to route between
MINIMALandHIGH, and map Pi reasoning levels to the model's supported thinking levelsFixed Gemini 2.5 Flash Lite minimal thinking budget to use the model's supported 512-token minimum instead of the regular Flash 128-token minimum, avoiding invalid thinking budget errors
Fixed OpenAI Codex Responses requests to forward configured
serviceTiervalues, restoring service-tier selection for Codex sessionsFixed
Container.render()stack overflow on long sessions by replacingArray.push(...spread)with a loop-based push, preventingRangeError: Maximum call stack size exceededwhen child output exceeds the V8 call stack argument limitFixed editor sticky-column tracking around paste markers so vertical cursor navigation restores the column from before the cursor entered a paste marker instead of jumping inside or past pasted content
Fixed queued messages typed during
/treebranch summarization to flush automatically after navigation completes, so they no longer remain stuck in the steering queue
0.67.0
13 Apr 2026Pi needs a reliable per-version usage signal to understand whether releases are being adopted and to help justify funding continued development.
Npm download counts are not a reliable proxy for actual Pi usage.
It only runs in interactive mode.
Telemetry
Interactive mode now sends a lightweight anonymous install/update telemetry ping to
https://pi.dev/install?version=x.y.zafter it writeslastChangelogVersioninsettings.json.Why this exists:
Pi needs a reliable per-version usage signal to understand whether releases are being adopted and to help justify funding continued development.
Npm download counts are not a reliable proxy for actual Pi usage.
How it works:
It only runs in interactive mode.
It does not run in RPC mode, print mode, JSON mode, or SDK mode.
On a fresh interactive install, Pi writes
lastChangelogVersion, then sends the ping.On later interactive startups, if the local changelog contains entries newer than the previously stored
lastChangelogVersion, Pi writes the newlastChangelogVersion, then sends the ping.The request is fire-and-forget. Startup does not wait for it, and any errors are ignored.
What data is collected:
Only the Pi version in the request path, for example
https://pi.dev/install?version=0.66.1.The server stores only aggregate per-version counters such as
{ "0.66.1": 3 }.It does not store IP addresses, client identifiers, prompts, paths, models, auth state, or any other per-user data. It literally only increments a counter for that version.
How to disable it:
/settings→ disableInstall telemetrysettings.json→ setenableInstallTelemetrytofalsePI_OFFLINE=1PI_TELEMETRY=0
New Features
Full
openRouterRoutingsupport inmodels.json, including fallbacks, parameter requirements, data collection, ZDR, ignore lists, quantizations, provider sorting, max price, and preferred throughput and latency constraints. See docs/models.md.PI_CODING_AGENT=trueenvironment variable set at startup so subprocesses can detect they are running inside the coding agent.
Added
Added full
openRouterRoutingfield support inmodels.json, including fallbacks, parameter requirements, data collection, ZDR, ignore lists, quantizations, provider sorting, max price, and preferred throughput and latency constraintsSet
PI_CODING_AGENT=trueenvironment variable at startup so sub-processes can detect they are running inside the coding agent
Fixed
Updated
antigravity-image-gen.tsexample extension to use User-Agent version1.21.9Bumped default Antigravity User-Agent version to
1.21.9Fixed Gemma 4 thinking level mapping to route between
MINIMALandHIGH, and map Pi reasoning levels to the model's supported thinking levelsFixed Gemini 2.5 Flash Lite minimal thinking budget to use the model's supported 512-token minimum instead of the regular Flash 128-token minimum, avoiding invalid thinking budget errors
Fixed OpenAI Codex Responses requests to forward configured
serviceTiervalues, restoring service-tier selection for Codex sessionsFixed
Container.render()stack overflow on long sessions by replacingArray.push(...spread)with a loop-based push, preventingRangeError: Maximum call stack size exceededwhen child output exceeds the V8 call stack argument limitFixed editor sticky-column tracking around paste markers so vertical cursor navigation restores the column from before the cursor entered a paste marker instead of jumping inside or past pasted content
Fixed queued messages typed during
/treebranch summarization to flush automatically after navigation completes, so they no longer remain stuck in the steering queue
- 0.66.x
0.66.1
8 Apr 2026Changed the Earendil announcement from an automatic startup notice to the hidden
/dementedelvesslash command.
0.66.0
8 Apr 2026Earendil startup announcement with bundled inline image rendering and a linked blog post for April 8 and 9, 2026.
Interactive Anthropic subscription auth warning when Anthropic subscription auth is active, clarifying that Anthropic third-party usage draws from extra usage and is billed per token.
Fixed bare
readlineimport to usenode:readlineprefix for Deno compatibility
New Features
Earendil startup announcement with bundled inline image rendering and a linked blog post for April 8 and 9, 2026.
Interactive Anthropic subscription auth warning when Anthropic subscription auth is active, clarifying that Anthropic third-party usage draws from extra usage and is billed per token.
Fixed
Fixed bare
readlineimport to usenode:readlineprefix for Deno compatibilityFixed auto-retry to treat stream failures like
request ended without sending any chunksas transient errorsFixed interactive startup notices to render after the initial resource listing, and added a bundled Earendil startup announcement with inline image rendering for April 8 and 9, 2026. Moved the blog link above the image to avoid overlap with terminal image rendering.
Fixed interactive mode to warn when Anthropic subscription auth is active, so users know Anthropic third-party usage draws from extra usage and is billed per token.
- 0.65.x
0.65.2
6 Apr 20260.65.1
5 Apr 2026Fixed bash output truncation by line count to always persist full output to a temp file, preventing data loss when output exceeds 2000 lines but stays under the byte threshold
RpcClient now forwards subprocess stderr to parent process in real-time
Theme file watcher now handles async
fs.watcherror events instead of crashing the process
Fixed
Fixed bash output truncation by line count to always persist full output to a temp file, preventing data loss when output exceeds 2000 lines but stays under the byte threshold
RpcClient now forwards subprocess stderr to parent process in real-time
Theme file watcher now handles async
fs.watcherror events instead of crashing the processFixed stored session cwd handling so resuming or importing a session whose original working directory no longer exists now prompts interactive users to continue in the current cwd, while non-interactive modes fail with a clear error.
Fixed resource collision precedence so project and user skills, prompt templates, and themes override package resources consistently, and CLI-provided paths take precedence over discovered resources
Fixed OpenAI-compatible completions streaming usage accounting to preserve
prompt_tokens_details.cache_write_tokensand normalize OpenRoutercached_tokens, preventing incorrect cache read/write token and cost reporting in piFixed CLI extension paths like
git:gist.github.com/...being incorrectly resolved against cwd instead of being passed through to the package managerFixed piped stdin runs with
--mode jsonto preserve JSONL output instead of falling back to plain textFixed interactive command docs to stop listing removed
/exitas a supported quit command
0.65.0
3 Apr 2026- Breaking
Removed extension post-transition events
session_switchandsession_fork. Usesession_startwithevent.reason("startup" | "reload" | "new" | "resume" | "fork"). For"new","resume", and"fork",session_startincludespreviousSessionFile. - Breaking
Removed session-replacement methods from
AgentSession. UseAgentSessionRuntimefornewSession(),switchSession(),fork(), andimportFromJsonl(). Cross-cwd session replacement rebuilds all cwd-bound runtime state and replaces the liveAgentSessioninstance. - Breaking
Removed
session_directoryfrom extension and settings APIs.
New Features
Session runtime API:
createAgentSessionRuntime()andAgentSessionRuntimeprovide a closure-based runtime that recreates cwd-bound services and session config on every session switch. Startup,/new,/resume,/fork, and import all use the same creation path. See docs/sdk.md and examples/sdk/13-session-runtime.ts.Label timestamps in
/tree: Toggle timestamps on tree entries withShift+T, with smart date formatting and timestamp preservation through branchingdefineTool()helper: Create standalone custom tool definitions with full TypeScript parameter type inference, no manual casts needed (#2746). See docs/extensions.md.Unified diagnostics: Arg parsing, service creation, session option resolution, and resource loading all return structured diagnostics (
info/warning/error) instead of logging or exiting. The app layer decides presentation and exit behavior.
Breaking Changes
- Breaking
Removed extension post-transition events
session_switchandsession_fork. Usesession_startwithevent.reason("startup" | "reload" | "new" | "resume" | "fork"). For"new","resume", and"fork",session_startincludespreviousSessionFile. - Breaking
Removed session-replacement methods from
AgentSession. UseAgentSessionRuntimefornewSession(),switchSession(),fork(), andimportFromJsonl(). Cross-cwd session replacement rebuilds all cwd-bound runtime state and replaces the liveAgentSessioninstance. - Breaking
Removed
session_directoryfrom extension and settings APIs. - Breaking
Unknown single-dash CLI flags (e.g.
-s) now produce an error instead of being silently ignored.
Migration: Extensions
Before:
pi.on("session_switch", async (event, ctx) => { ... }); pi.on("session_fork", async (_event, ctx) => { ... });After:
pi.on("session_start", async (event, ctx) => { // event.reason: "startup" | "reload" | "new" | "resume" | "fork" // event.previousSessionFile: set for "new", "resume", "fork" });Migration: SDK session replacement
Before:
await session.newSession(); await session.switchSession("/path/to/session.jsonl");After:
import { type CreateAgentSessionRuntimeFactory, createAgentSessionFromServices, createAgentSessionRuntime, createAgentSessionServices, getAgentDir, SessionManager, } from "@mariozechner/pi-coding-agent"; const createRuntime: CreateAgentSessionRuntimeFactory = async ({ cwd, sessionManager, sessionStartEvent }) => { const services = await createAgentSessionServices({ cwd }); return { ...(await createAgentSessionFromServices({ services, sessionManager, sessionStartEvent })), services, diagnostics: services.diagnostics, }; }; const runtime = await createAgentSessionRuntime(createRuntime, { cwd: process.cwd(), agentDir: getAgentDir(), sessionManager: SessionManager.create(process.cwd()), }); await runtime.newSession(); await runtime.switchSession("/path/to/session.jsonl"); await runtime.fork("entry-id"); // After replacement, runtime.session is the new live session. // Rebind any session-local subscriptions or extension bindings.Added
Added
createAgentSessionRuntime()andAgentSessionRuntimefor runtime-backed session replacement. The runtime takes aCreateAgentSessionRuntimeFactoryclosure that closes over process-global fixed inputs and recreates cwd-bound services and session config for each effective cwd. Startup and later/new,/resume,/fork, import all use the same factory.Added unified diagnostics model (
info/warning/error) for arg parsing, service creation, session option resolution, and resource loading. Creation logic no longer logs or exits. The app layer decides presentation and exit behavior.Added error diagnostics for missing explicit CLI resource paths (
-e,--skill,--prompt-template,--theme)Added
defineTool()so standalone and array-based custom tool definitions keep inferred parameter types without manual castsAdded label timestamps to the session tree with a
Shift+Ttoggle in/tree, smart date formatting, and timestamp preservation through branching
Fixed
Fixed startup resource loading to reuse the initial
ResourceLoaderfor the first runtime, so extensions are not loaded twice before session startup andsession_starthandlers still fire for singleton-style extensionsFixed retry settlement so retried agent runs wait for the full retry cycle to complete before declaring idle, preventing stale state after transient errors
Fixed theme
exportcolors to resolve theme variables the same way ascolors, so/exportHTML backgrounds now honor entries likepageBg: "base"instead of requiring inline hex valuesFixed Bedrock throttling errors being misidentified as context overflow, causing unnecessary compaction instead of retry
Added tool streaming support for newer Z.ai models
- 0.64.x
0.64.0
29 Mar 2026- Breaking
ModelRegistryno longer has a public constructor. SDK callers and tests must useModelRegistry.create(authStorage, modelsJsonPath?)for file-backed registries orModelRegistry.inMemory(authStorage)for built-in-only registries. Directnew ModelRegistry(...)calls no longer compile. Extensions and SDK callers can attach a
prepareArgumentshook to any tool definition, letting them normalize or migrate raw model arguments before schema validation. The built-inedittool uses this to transparently support sessions created with the old single-edit schema. See docs/extensions.mdExtensions can customize the collapsed thinking block label via
ctx.ui.setHiddenThinkingLabel(). See examples/extensions/hidden-thinking-label.ts
New Features
Extensions and SDK callers can attach a
prepareArgumentshook to any tool definition, letting them normalize or migrate raw model arguments before schema validation. The built-inedittool uses this to transparently support sessions created with the old single-edit schema. See docs/extensions.mdExtensions can customize the collapsed thinking block label via
ctx.ui.setHiddenThinkingLabel(). See examples/extensions/hidden-thinking-label.ts
Breaking Changes
- Breaking
ModelRegistryno longer has a public constructor. SDK callers and tests must useModelRegistry.create(authStorage, modelsJsonPath?)for file-backed registries orModelRegistry.inMemory(authStorage)for built-in-only registries. Directnew ModelRegistry(...)calls no longer compile.
Added
Added
ToolDefinition.prepareArgumentshook to prepare raw tool call arguments before schema validation, enabling compatibility shims for resumed sessions with outdated tool schemasBuilt-in
edittool now usesprepareArgumentsto silently fold legacy top-leveloldText/newTextintoeditswhen resuming old sessionsAdded
ctx.ui.setHiddenThinkingLabel()so extensions can customize the collapsed thinking label in interactive mode, with a no-op in RPC mode and a runnable example extension inexamples/extensions/hidden-thinking-label.ts
Fixed
Fixed extension-queued user messages to refresh the interactive pending-message list so messages submitted while a turn is active are no longer silently dropped
Fixed monorepo
tsconfig.jsonpath mappings to resolve@mariozechner/pi-aisubpath exports to source files in development checkoutsFixed TUI cell size response handling to consume only exact
CSI 6 ; height ; width treplies, so bareEscapeis no longer swallowed while waiting for terminal image metadataFixed Kitty keyboard protocol keypad functional keys to normalize to logical digits, symbols, and navigation keys, so numpad input in terminals such as iTerm2 no longer inserts Private Use Area gibberish or gets ignored
- 0.63.x
0.63.2
29 Mar 2026Extension handlers can now use
ctx.signalto forward cancellation into nested model calls,fetch(), and other abort-aware work. See docs/extensions.md#ctxsignalBuilt-in
edittool input now useseditsas the only replacement shape, reducing invalid tool calls caused by mixed single-edit and multi-edit schemasLarge multi-edit results no longer trigger full-screen redraws in the interactive TUI when the final diff is rendered
New Features
Extension handlers can now use
ctx.signalto forward cancellation into nested model calls,fetch(), and other abort-aware work. See docs/extensions.md#ctxsignalBuilt-in
edittool input now useseditsas the only replacement shape, reducing invalid tool calls caused by mixed single-edit and multi-edit schemasLarge multi-edit results no longer trigger full-screen redraws in the interactive TUI when the final diff is rendered
Added
Added
ctx.signaltoExtensionContextand wired it to the active agent turn so extension handlers can forward cancellation into nested model calls,fetch(), and other abort-aware work
Fixed
Fixed built-in
edittool input to useeditsas the only replacement shape, eliminating the mixed single-edit and multi-edit modes that caused repeated invalid tool calls and retriesFixed edit tool TUI rendering to defer large multi-edit diffs to the settled result, avoiding full-screen redraws when the tool completes
0.63.1
27 Mar 2026Added
gemini-3.1-pro-preview-customtoolsmodel availability for thegoogle-vertexproviderDocumented
tool_callinput mutation as supported extension API behavior, clarified that post-mutation inputs are not re-validated, and added regression coverage for executing mutated tool argumentsFixed repeated compactions dropping messages that were kept by an earlier compaction by re-summarizing from the previous kept boundary and recalculating
tokensBeforefrom the rebuilt session context
Added
Added
gemini-3.1-pro-preview-customtoolsmodel availability for thegoogle-vertexprovider
Fixed
Documented
tool_callinput mutation as supported extension API behavior, clarified that post-mutation inputs are not re-validated, and added regression coverage for executing mutated tool argumentsFixed repeated compactions dropping messages that were kept by an earlier compaction by re-summarizing from the previous kept boundary and recalculating
tokensBeforefrom the rebuilt session contextFixed interactive compaction UI updates so
ctx.compact()rebuilds the chat through unified compaction events, manual compaction no longer duplicates the summary block, and thetrigger-compactexample only fires when context usage crosses its thresholdFixed interactive compaction completion to append a synthetic compaction summary after rebuilding the chat so the latest compaction remains visible at the bottom
Fixed skill discovery to stop recursing once a directory contains
SKILL.md, and to ignore root*.mdfiles in.agents/skillswhile keeping root markdown skill files supported in~/.pi/agent/skills,.pi/skills, and packageskills/directoriesFixed edit tool diff rendering for multi-edit operations with large unchanged gaps so distant edits collapse intermediate context instead of dumping the full unchanged middle block
Fixed edit tool error rendering to avoid repeating the same exact-match failure in both the preview and result blocks
Fixed auto-compaction overflow recovery for Ollama models when the backend returns explicit
prompt too long; exceeded max context length ...errors instead of silently truncating inputFixed built-in tool overrides that reuse built-in parameter schemas to still honor custom
renderCallandrenderResultrenderers in the interactive TUI, restoring theminimal-modeexample
0.63.0
27 Mar 2026- Breaking
ModelRegistry.getApiKey(model)has been replaced bygetApiKeyAndHeaders(model)becausemodels.jsonauth and header values can now resolve dynamically on every request. Extensions and SDK integrations that previously fetched only an API key must now fetch request auth per call and forward bothapiKeyandheaders. UsegetApiKeyForProvider(provider)only when you explicitly want provider-level API key lookup without model headers orauthHeaderhandling - Breaking
Removed deprecated direct
minimaxandminimax-cnmodel IDs, keeping onlyMiniMax-M2.7andMiniMax-M2.7-highspeed. Update pinned model IDs to one of those supported direct MiniMax models, or use another provider route that still exposes the older IDs Added
sessionDirsetting support in global and projectsettings.jsonso session storage can be configured without passing--session-diron every invocation
Breaking Changes
- Breaking
ModelRegistry.getApiKey(model)has been replaced bygetApiKeyAndHeaders(model)becausemodels.jsonauth and header values can now resolve dynamically on every request. Extensions and SDK integrations that previously fetched only an API key must now fetch request auth per call and forward bothapiKeyandheaders. UsegetApiKeyForProvider(provider)only when you explicitly want provider-level API key lookup without model headers orauthHeaderhandling - Breaking
Removed deprecated direct
minimaxandminimax-cnmodel IDs, keeping onlyMiniMax-M2.7andMiniMax-M2.7-highspeed. Update pinned model IDs to one of those supported direct MiniMax models, or use another provider route that still exposes the older IDs
Migration Notes
Before:
const apiKey = await ctx.modelRegistry.getApiKey(model); return streamSimple(model, messages, { apiKey });After:
const auth = await ctx.modelRegistry.getApiKeyAndHeaders(model); if (!auth.ok) throw new Error(auth.error); return streamSimple(model, messages, { apiKey: auth.apiKey, headers: auth.headers, });Added
Added
sessionDirsetting support in global and projectsettings.jsonso session storage can be configured without passing--session-diron every invocationAdded a startup onboarding hint in the interactive header telling users pi can explain its own features and documentation
Added
edittool multi-edit support so one call can update multiple separate, disjoint regions in the same file while matching all replacements against the original file contentAdded support for
PI_TUI_WRITE_LOGdirectory paths, creating a unique log file (tui-<timestamp>-<pid>.log) per instance for easier debugging of multiple pi sessions
Fixed
Fixed file mutation queue ordering so concurrent
editandwriteoperations targeting the same file stay serialized in request order instead of being reordered during queue-key resolutionFixed
models.jsonshell-command auth and headers to resolve at request time instead of being cached into long-lived model state. pi now leaves TTL, caching, and recovery policy to user-provided wrapper commands because arbitrary shell commands need provider-specific strategiesFixed Google and Vertex cost calculation to subtract cached prompt tokens from billable input tokens instead of double-counting them when providers report
cachedContentTokenCountAdded missing
ajvdirect dependency; previously relied on transitive install via@mariozechner/pi-aiwhich broke standalone installsFixed
/exportHTML backgrounds to honortheme.export.pageBg,cardBg, andinfoBginstead of always deriving them fromuserMessageBgFixed interactive bash execution collapsed previews to recompute visual line wrapping at render time, so previews respect the current terminal width after resizes and split-pane width changes
Fixed RPC
get_session_statsto exposecontextUsage, so headless clients can read actual current context-window usage instead of deriving it from token totalsFixed
pi updatefor git packages to fetch only the tracked target branch with--no-tags, reducing unrelated branch and tag noise while preserving force-push-safe updatesFixed print and JSON modes to emit
session_shutdownbefore exit, so extensions can release long-lived resources and non-interactive runs terminate cleanlyFixed GitHub Copilot OpenAI Responses requests to omit the
reasoningfield entirely when no reasoning effort is requested, avoiding400errors from Copilotgpt-5-minirejectingreasoning: { effort: "none" }during internal summary callsFixed blockquote text color breaking after inline links (and other inline elements) due to missing style restoration prefix
Fixed slash-command Tab completion from immediately chaining into argument autocomplete after completing the command name, restoring flows like
/modelthat submit into a selector dialogFixed stale content and incorrect viewport tracking after TUI content shrinks or transient components inflate the working area
Fixed
@autocomplete to debounce editor-triggered searches, cancel in-flightfdlookups cleanly, and keep suggestions visible while results refresh
- 0.62.x
0.62.0
23 Mar 2026Reading the note…
- 0.61.x
0.61.1
20 Mar 2026Reading the note…
0.61.0
20 Mar 2026Reading the note…
- 0.60.x
0.60.0
18 Mar 2026Reading the note…
- 0.59.x
0.59.0
17 Mar 2026- Breaking
Changed custom tool system prompt behavior so extension and SDK tools are included in the default
Available toolssection only when they providepromptSnippet. OmittingpromptSnippetnow leaves the tool out of that section instead of falling back todescription Faster startup by lazy-loading
@mariozechner/pi-aiprovider SDKs on first use instead of import timeBetter provider retry behavior when providers return error messages as responses
New Features
Faster startup by lazy-loading
@mariozechner/pi-aiprovider SDKs on first use instead of import timeBetter provider retry behavior when providers return error messages as responses
Better terminal integration via OSC 133 command-executed markers
Better Git footer branch detection for repositories using reftable storage
Breaking Changes
- Breaking
Changed custom tool system prompt behavior so extension and SDK tools are included in the default
Available toolssection only when they providepromptSnippet. OmittingpromptSnippetnow leaves the tool out of that section instead of falling back todescription
Changed
Lazy-load built-in
@mariozechner/pi-aiprovider modules and root provider wrappers so coding-agent startup no longer eagerly loads provider SDKs before first use
Fixed
Fixed session title handling in
/tree, compaction, and branch summarization so empty title clears render correctly andsession_infoentries stay out of summariesFixed footer branch detection for Git repositories using reftable storage so branch names still appear correctly in the footer
Fixed rendered user messages to emit an OSC 133 command-executed marker after command output, improving terminal prompt integration
Fixed provider retry handling to treat provider-returned error messages as retryable failures instead of successful responses
Fixed Claude 4.6 context window overrides in bundled model metadata so coding-agent sees the intended model limits after generated catalogs are rebuilt
- 0.58.x
0.58.4
16 Mar 2026Fixed steering messages to wait until the current assistant message's tool-call batch fully finishes instead of skipping pending tool calls.
0.58.3
15 Mar 20260.58.2
15 Mar 2026Improved settings, theme, thinking, and show-images selector layouts by using configurable select-list primary column sizing
Fixed fuzzy
editmatching to normalize Unicode compatibility variants before comparison, reducing false "oldText not found" failures for text such as CJK and full-width charactersFixed
/model <ref>exact matching and picker search to recognize canonicalprovider/modelreferences when model IDs themselves contain/, such as LM Studio models likeunsloth/qwen3.5-35b-a3b
Added
Improved settings, theme, thinking, and show-images selector layouts by using configurable select-list primary column sizing
Fixed
Fixed fuzzy
editmatching to normalize Unicode compatibility variants before comparison, reducing false "oldText not found" failures for text such as CJK and full-width charactersFixed
/model <ref>exact matching and picker search to recognize canonicalprovider/modelreferences when model IDs themselves contain/, such as LM Studio models likeunsloth/qwen3.5-35b-a3bFixed Anthropic OAuth manual login and token refresh by using the localhost callback URI for pasted redirect/code flows and omitting
scopefrom refresh-token requestsFixed stale scrollback remaining after session switches by clearing the screen before wiping scrollback
Fixed extra blank lines after markdown block elements in rendered output
0.58.1
14 Mar 2026Added
pi uninstallalias forpi install --uninstallconvenienceFixed OpenAI Codex websocket protocol to include required headers and properly terminate SSE streams on connection close
Fixed WSL clipboard image fallback to properly handle missing clipboard utilities and permission errors
Added
Added
pi uninstallalias forpi install --uninstallconvenience
Fixed
Fixed OpenAI Codex websocket protocol to include required headers and properly terminate SSE streams on connection close
Fixed WSL clipboard image fallback to properly handle missing clipboard utilities and permission errors
Fixed extension
session_starthook firing before TUI was ready, causing UI operations insession_starthandlers to failFixed Windows shell and path handling for package manager operations and autocomplete to properly handle drive letters and mixed path separators
Fixed Bedrock prompt caching being enabled for non-Claude models, causing API errors
Fixed Qwen models via OpenAI-compatible providers by adding
qwen-chat-templatecompat mode that uses Qwen's native chat template formatFixed Bedrock unsigned thinking replay to handle edge cases with empty or malformed thinking blocks
Fixed headless clipboard fallback logging spurious errors in non-interactive environments
Fixed
models.jsonprovider compat flags not being honored when loading custom model definitionsFixed xhigh reasoning effort detection for Claude Opus 4.6 to match by model ID instead of requiring explicit capability flag
Fixed prompt cwd containing Windows backslashes breaking bash tool execution by normalizing to forward slashes
Fixed editor paste to preserve literal content instead of normalizing newlines, preventing content corruption for text with embedded escape sequences
Fixed skill discovery recursing past skill root directories when nested SKILL.md files exist
Fixed tab completion to preserve
./prefix when completing relative pathsFixed npm package installs and lookups being tied to the active repository Node version by adding
npmCommandas an argv-style settings override for package manager operationsFixed
ctx.ui.getEditorText()in the extension API returning paste markers (e.g.,[paste #1 +24 lines]) instead of the actual pasted contentFixed startup crash when downloading
fd/ripgrepon first run by usingpipeline()instead offinished(readable.pipe(writable))so stream errors from timeouts are caught properly, and increased the download timeout from 10s to 120s
0.58.0
14 Mar 2026Claude Opus 4.6, Sonnet 4.6, and related Bedrock models now use a 1M token context window (up from 200K)
Extension tool calls now execute in parallel by default, with sequential
tool_callpreflight preserved for extension interception.GOOGLE_CLOUD_API_KEYenvironment variable support for thegoogle-vertexprovider as an alternative to Application Default Credentials
New Features
Claude Opus 4.6, Sonnet 4.6, and related Bedrock models now use a 1M token context window (up from 200K)
Extension tool calls now execute in parallel by default, with sequential
tool_callpreflight preserved for extension interception.GOOGLE_CLOUD_API_KEYenvironment variable support for thegoogle-vertexprovider as an alternative to Application Default CredentialsExtensions can supply deterministic session IDs via
newSession()
Added
Added
GOOGLE_CLOUD_API_KEYenvironment variable support for thegoogle-vertexprovider as an alternative to Application Default CredentialsAdded custom session ID support in
newSession()for extensions that need deterministic session paths
Changed
Changed extension tool interception to use agent-core
beforeToolCallandafterToolCallhooks instead of wrapper-based interception. Tool calls now execute in parallel by default, extensiontool_callpreflight still runs sequentially, and final tool results are emitted in assistant source order.Raised Claude Opus 4.6, Sonnet 4.6, and related Bedrock model context windows from 200K to 1M tokens
Fixed
Fixed
tool_callextension handlers observing stalesessionManagerstate during multi-tool turns by draining queued agent events before eachtool_callpreflight. In parallel tool mode this guarantees state through the current assistant tool-calling message, but not sibling tool results from the same assistant message.Fixed interactive input fields backed by the TUI
Inputcomponent to scroll by visual column width for wide Unicode text (CJK, fullwidth characters), preventing rendered line overflow and TUI crashes in places like search and filter inputsFixed
shift+taband other modified Tab bindings in tmux whenextended-keys-formatis left at the defaultxtermFixed EXIF orientation not being applied during image convert and resize, causing JPEG and WebP images from phone cameras to display rotated or mirrored
Fixed the default coding-agent system prompt to include only the current date in ISO format, not the current time, so prompt prefixes stay cacheable across reloads and resumed sessions
Fixed retry regex to match
server_errorandinternal_errorerror types from providers, improving automatic retry coverageFixed example extensions to support
PI_CODING_AGENT_DIRenvironment variable for custom agent directory pathsFixed tool result images not being sent in
function_call_outputitems for OpenAI Responses API providers, causing image data to be silently dropped in tool resultsFixed assistant content being sent as structured content blocks instead of plain strings in the
openai-completionsprovider, causing errors with some OpenAI-compatible backendsFixed error details in OpenAI Responses
response.failedhandler to include status code, error code, and message instead of a generic failureFixed GitHub Copilot device-code login polling to respect OAuth slow-down intervals, wait before the first token poll, and include a clearer clock-drift hint in WSL/VM environments when repeated slow-downs lead to timeout
Fixed usage statistics not being captured for OpenAI-compatible providers that return usage in
choice.usageinstead of the standardchunk.usage(e.g., Moonshot/Kimi)Fixed editor scroll indicator rendering crash in narrow terminal widths
Fixed tab characters in editor and input paste not being normalized to spaces (#2027, #1975
Fixed
wordWrapLineoverflow when wide characters (CJK, fullwidth) fall exactly at the wrap boundaryFixed paste markers not being treated as atomic segments in editor word wrapping and cursor navigation
- 0.57.x
0.57.1
7 Mar 2026Tree branch folding and segment-jump navigation in
/tree, withCtrl+←/Ctrl+→andAlt+←/Alt+→shortcuts while←/→andPage Up/Page Downremain available for paging. See docs/tree.md and docs/keybindings.md.session_directoryextension event for customizing session directory paths before session manager creation. See docs/extensions.md.Digit keybindings (
0-9) in the TUI keybinding system, including modified combos likectrl+1. See docs/keybindings.md.
New Features
Tree branch folding and segment-jump navigation in
/tree, withCtrl+←/Ctrl+→andAlt+←/Alt+→shortcuts while←/→andPage Up/Page Downremain available for paging. See docs/tree.md and docs/keybindings.md.session_directoryextension event for customizing session directory paths before session manager creation. See docs/extensions.md.Digit keybindings (
0-9) in the TUI keybinding system, including modified combos likectrl+1. See docs/keybindings.md.
Added
Added
/treebranch folding and segment-jump navigation withCtrl+←/Ctrl+→andAlt+←/Alt+→, while keeping←/→andPage Up/Page Downfor pagingAdded
session_directoryextension event that fires before session manager creation, allowing extensions to customize the session directory path based on cwd and other factors. CLI--session-dirflag takes precedence over extension-provided pathsAdded digit keys (
0-9) to the keybinding system, including Kitty CSI-u and xtermmodifyOtherKeyssupport for bindings likectrl+1
Fixed
Fixed custom tool collapsed/expanded rendering in HTML exports. Custom tools that define different collapsed vs expanded displays now render correctly in exported HTML, with expandable sections when both states differ and direct display when only expanded exists
Fixed tmux startup guidance and keyboard setup warnings for modified key handling, including Ghostty
shift+enter=text:\nremap guidance and tmuxextended-keys-formatdetectionFixed z.ai context overflow recovery so
model_context_window_exceedederrors trigger auto-compaction instead of surfacing as unhandled stop reason failuresFixed autocomplete selection ignoring typed text: highlight now follows the first prefix match as the user types, and exact matches are always selected on Enter
Fixed slash-command Tab completion to immediately open argument completions when available
Fixed explicit
pi -e <path>extensions losing command and tool conflicts to discovered extensions by giving CLI-loaded extensions higher precedenceFixed Windows external editor launch for
Ctrl+Gandctx.ui.editor()so shell-based commands likeEDITOR="code --wait"work correctly
0.57.0
7 Mar 2026Reading the note…
- 0.56.x
0.56.3
6 Mar 2026Reading the note…
0.56.2
5 Mar 2026Reading the note…
0.56.1
5 Mar 2026Reading the note…
0.56.0
4 Mar 2026Reading the note…
A whole project to check? Paste its package.json in the dependency check: every deprecated, outdated or vulnerable dependency, with what to upgrade to. Free, no account.