v2026.6.17-2

Added

Fixed

• Fixed reusable tool renderer components to survive repeated TUI refreshes after long-session cleanup changes.

Changed

Removed

v2026.6.17

Added

• Added provider-scoped auth.json environment overrides and a global httpProxy setting inherited from upstream.

Fixed

• Fixed HTTP dispatcher configuration to preserve caller-provided fetch overrides.
• Fixed inherited OpenAI Responses null-content handling and DeepSeek thinking-off compatibility.
• Fixed device-code login to avoid opening the browser automatically.

Changed

• Updated the vendored Markdown parser used by HTML session exports to marked 18.0.5.

Removed

v2026.6.16-2

Added

Fixed

Changed

Removed

v2026.6.16

Added

• Added a compaction trim benchmark suite for hard-limit emergency pruning.

Fixed

• Fixed inherited Google and google-vertex Gemini model metadata to map latest aliases to the current models, add Gemini 3.5 Flash for Vertex, correct Gemini 2.5 Flash Vertex cache pricing, and remove shut-down Vertex preview models (#5761).
• Fixed the session selector to stay open and show the all-sessions empty state when both current-folder and all-scope session lists are empty (#5747).
• Fixed inherited Moonshot AI China model metadata to include Kimi K2.7 Code, and omitted unsupported thinking-off payloads for Kimi K2.7 Code models (#5760).
• Compaction now keeps the last valid cut point when the recent-token budget is exceeded entirely within trailing tool results, instead of silently retaining the whole history.

Changed

• Improved hard-limit emergency pruning so old-message trimming tracks a running token total instead of re-estimating the whole conversation each iteration.
• Improved extension context emission by using a JSON-only clone for the hot context message copy.
• Improved single-line intra-line diff rendering with conservative fast paths for identical, long, and single-span edits.

Removed

v2026.6.15

Added

Fixed

Changed

Removed

v2026.6.14-3

Added

Fixed

• Fixed interactive startup rendering so the initial TUI screen is not duplicated.

Changed

Removed

v2026.6.14-2

Added

Fixed

• Fixed startup so a broken extension reports a warning instead of aborting the entire senpi process.

Changed

Removed

v2026.6.14

Added

• Added first-run interactive theme detection from the terminal background.

Fixed

• Fixed pi update for pnpm global installs whose configured global-bin-dir no longer matches the active pnpm home (#5689).
• Fixed npm package specs that use ranges or tags (for example @^1.2.7) so installed package resources still load instead of being treated as mismatched exact pins (#5695).
• Fixed inherited OpenCode/OpenCode Go completion model metadata to omit long-retention cache fields for routes that reject prompt_cache_retention (#5702).
• Fixed custom provider config so plain uppercase API key and header values remain literals instead of being treated as legacy environment references; use explicit $ENV_VAR syntax for environment variables (#5661).

Changed

Removed

v2026.6.13

Added

• Added JSONL parse and RPC event emission benchmark harnesses.
• Added an experimental first-time setup flow behind PI_EXPERIMENTAL=1 that asks for a dark/light theme choice (preselecting the detected appearance) and opt-in analytics data sharing on first launch with the default agent directory; opting in stores a trackingId in settings.json (#5587 by @vegarsti).
• Added AWS data retention documentation links to inherited Amazon Bedrock unsupported data retention mode validation errors (#5561 by @unexge).

Fixed

• Fixed project trust detection to ignore global ~/.senpi/agent state when running from $HOME, and made senpi update use only saved or explicit project trust without prompting (#5619).
• Fixed experimental first-time setup to skip forked sessions instead of rerunning the setup prompts (#5627 by @vegarsti).
• Fixed inherited OpenAI-compatible context overflow detection for parenthesized maximum context length (N) errors (#5677).
• Fixed inherited OpenAI GPT-5.4/GPT-5.5 and OpenAI Codex GPT-5.4/GPT-5.4 mini/GPT-5.5 context window metadata to use the observed 272k-token Codex backend limit, avoiding a billing hazard from prompts above Codex's accepted limit (reported by @trethore).
• Fixed inherited OpenAI GPT-5.4/GPT-5.5 and OpenAI Codex GPT-5.4/GPT-5.4 mini/GPT-5.5 context window metadata to match current OpenAI limits (#5644).
• Fixed inherited Anthropic refusal stops to preserve provider stop_details explanations in error messages (#5666 by @rwachtler).
• Increased the inherited OpenAI Codex Responses SSE response-header timeout to 20 seconds to reduce false-positive stalls while retaining the bounded wait introduced for zero-event hangs (#4945).
• Fixed inherited Claude Fable 5 thinking-off requests to omit Anthropic's unsupported thinking.type: "disabled" payload (#5567 by @tmustier).
• Fixed inherited late tool progress callbacks after tool settlement to be ignored instead of emitting stale tool_execution_update events (#5573).
• Fixed inherited user-message transcript rendering so standalone + messages no longer render as - (#5657).
• Fixed inherited slash-separated fuzzy queries so provider/model completions remain matchable after insertion.
• Fixed inherited WezTerm inline Kitty image rendering so reserved row clears do not erase all but the top strip of tool image previews (#5618).
• Fixed inherited editor wrapping for CJK text to break at character boundaries instead of leaving large trailing gaps (#5585 by @haoqixu).
• Fixed inherited loose Markdown list rendering to preserve blank-line separation between list items (#5562 by @Perlence).
• Fixed --model resolution for authenticated custom model IDs whose slash prefix matches an unauthenticated built-in provider (#5643).
• Fixed /fork to keep session parent chains connected when the forked path contains labels (#5669).
• Fixed /share and /export HTML exports to use the active fallback theme when the configured custom theme no longer exists (#5596).
• Fixed custom fallback model IDs with :<thinking> suffixes to preserve the requested thinking level when the provider template model does not advertise reasoning (#5560 by @haoqixu).
• Fixed resumed aborted tool calls to reuse the persisted retry-attempt label.
• Fixed Bun binaries ignoring project .env provider credentials so untrusted project files cannot inject API keys.
• Fixed print mode to print preceding assistant text when a terminating tool result is the final message; empty assistant text still prints nothing.

Changed

• Improved RPC event emission by coalescing same-tick event frames into one stdout write, reducing the benchmarked write count by 99.90%.
• Reduced Bun compiled binary size by 7.06% with minification while preserving function names.
• Models registry Anthropic compatibility config now accepts supportsToolChoice and supportsForcedToolChoice flags.

Removed

v2026.6.12-2

Added

Fixed

Removed