Untested on omp
-
A TLH-style git status add-on for pi's built-in footer.
-
Pi extension package for web search and URL-native document fetching
-
Ask the user questions via a Pi tool
-
Cloudflare AI Gateway Smart Router support and routing diagnostics for Pi
-
Exact Catppuccin themes for Pi, plus optional Catppuccin TUI polish.
-
A pi extension extension that brings smart Claude Code-style autonomous sub-agents to pi.
-
pi extension: MCP Tasks Extension (SEP-2663) client over pi's registered MCP servers
-
Peak/off-peak billing-hours + promotion indicator for the pi coding agent. Config-driven: providers, models, and time-windowed campaigns via a JSON data file. Append-only footer status, /peak all-models report.
-
Verbatim Jev-guided context compaction for pi: replaces the built-in compaction summary with fast-jev-compaction decisions — stale tool calls/results are dropped or truncated, everything kept stays verbatim.
-
pi 本地配置管理:Provider、模型、独立 API Key 与默认设置,支持 pi 扩展和独立启动
-
TypeSafe Jev (System One judgments) for Pi: zero-dependency jev_ask tool with question linting, model discovery, probability-bar rendering, and daily budget caps.
-
Automatic isolated secondary review for Pi coding agent sessions
-
Herdr-native side-agent orchestration for Pi with one Git worktree per pane
-
Rounded frames around assistant reply bodies in pi — thinking blocks stay unframed.
-
Opt-in Lune appearance for Pi: pixel-moon header, seven moonlit palettes and Lune tool cards
-
把《豆包工作蓝皮书》(AlephAITech/DoubaoWorkGuide,MIT,50 篇正文)打包成本地 Agent Skill:SKILL.md 只做意图路由,正文按需加载,附增量同步脚本。
-
Pi 复合工程工作流引擎(状态感知 / 断点续跑 / 输出压缩 / 自动化工具集)
-
Standalone Pi Mentis memory extension for Pi >= 0.84.0
-
Theme-aware netrunner flavor for the pi coding agent: Night City loader lines, glitch effects, spinners, NetWatch trace, a glitch logo banner and an optional neon theme
-
Google Vertex AI provider for Pi coding agent with dynamic model discovery
-
Pi extension for any local model harness - model discovery, per-project config
-
Context-efficient web search and fetch tools for Pi.
-
Pi extension that rewrites the agent's bash to the modern toolchain: grep/egrep/fgrep → rg (ripgrep), find → fd (conservative whitelist) — applied at the tool_call event, so it composes with other bash extensions
-
Run pi sandboxed in a QEMU microVM via Gondolin
-
Vim-style transcript selection, copying, annotations, and reactions for Pi
-
Opt-in hold-Space Deepgram dictation for Pi
-
Bridge agy CLI Gemini models into Pi via the agy CLI tool
-
Pi extension that shows a token-budget breakdown of the assembled system prompt
-
Live tokens-per-second (TPS) meter for Pi with gentle-pi subagent visibility
-
Pi extension that keeps supported provider prompt caches warm during long idle gaps
-
NVIDIA NIM provider for pi — full model catalog, per-family thinking handlers, and aggressive rate-limit retry
-
Prompt workflow extension for Pi coding agent
-
Prevent macOS idle sleep while Pi agents are working.
-
Read-only right sidebar for the Thoth Pi ecosystem.
-
Hard budget enforcement for pi: warn, auto-downgrade to a cheaper model, and block requests before you overspend.
-
D-Mail — model-driven context folding for Pi.
-
Manage your todos through Notion: keeps a local SQLite mirror of your Notion task databases for fast reads, time tracking, and agent-driven task management
-
Mix, preview, and apply reusable prompt snippets in Pi.
-
Minimal footer with model, usage, and project details for Pi.
-
Same-model Fast / Priority scheduling for Pi on OpenAI and xAI.
-
Persistent scheduled Pi jobs backed by a detached daemon and macOS launchd.
-
Structured planning, TDD, and Git worktrees for Pi local coding agents
-
Durable agent memory for the Pi coding agent: every completed turn becomes an owner-controlled Markdown episode, with exact, ranked, bounded recall.
-
a simple, standalone webui for pi.dev
-
Pi extension for a highly customizable, ANSI-aware statusline footer.
-
Isolated task delegation, Markdown roles, and session-local background agents for Pi.
-
Engineering research agent: discover, read, synthesize, verify, and review engineering knowledge across mechanical, software, civil, electrical, and architectural disciplines.
-
Problem-solving canvas toolface for pi: six deferred canvas_* tools bridging the agent to the taiji runtime CanvasService via the CANVAS_MARKER select channel (problem-solve-canvas)
-
Portable Pi skills for maintained code
-
Smart per-turn model router for Pi that optimizes your AI budget and usage limits by dynamically switching LLM tiers based on task complexity, budget, and phase awareness.