Untested on omp

Make pi's Anthropic OAuth requests match the current Claude Code client so OAuth login works correctly.

Run [Berget AI](https://berget.ai) models inside [Pi](https://pi.dev).

Persistent memory for pi — learns corrections, preferences, and patterns from sessions and injects them into future conversations.

Shared local memory for coding agents with semantic recall, knowledge graph, and code navigation.

Personal extensions for the [Pi coding agent](https://github.com/badlogic/pi-mono).

A single, customizable Pi UI package with layout-oriented Context and Shell rendering for Pi.

Render fenced code blocks in Pi's transcript as styled, flush-edge cards with one-click copy and a raw-text toggle.

Native Pi extension and lookup skill for evidence-grounded IETF RFC research.

Patch provider system prompts with exact, config-driven replacements.

StepFun provider for pi — Step 5 Preview and Flash via pay-as-you-go or Step Plan

Keyless OpenCode Zen free models for Pi, using native HTTP APIs

Define arbitrary mention sources and triggers for Pi.

Session-native Notes, History, and agent-owned Context Windows for Pi.

Search the web, extract pages, get grounded answers, and run deep research with the web CLI, a TypeScript library, and a pi extension. Bring your own providers and API keys.

Tern integration for the pi coding agent: TSP handshake, native Mermaid diagrams, WKWebView browser automation, pane capture and a session mirror.

Asynchronous subagent delegation for pi with run, batch, chain, and continuation workflows.

Goal mode extension for pi: durable long-running objectives, a constant six-tool surface, structured tasks, independent completion audit, Sisyphus mode, auto-continue, and a status overlay.

Path Guard for pi — blocks destructive commands & path overwrites, protects .env/keys, with strict/normal/loose/trusted/naked guard modes (/guard). pi 防误删/防误覆盖扩展，支持 5 种防护模式。

Background review and attributed corrective advice for Pi agent sessions.

pi extension that adds Fast and Ultrafast service tiers to supported OpenAI requests.

Configurable Jev decision layer for OmO and senpi

Pi extension: discover models from {baseUrl}/models for providers marked with discoverModels in models.json

Token-efficient MCP adapter for the Pi coding agent

Hashline read/edit/grep tool overrides for Pi

Pi extension that provides native Munin long-term memory tools through direct SDK integration.

Pi extension to keep context usage near a configurable soft limit

WebFetch tool extension for pi coding agent

Interactive decisions, review triage, and human handoffs for Pi.

Publish markdown, CSV and Excel files to Google Drive as native Google Docs and Sheets, rewriting relative links to real Doc URLs with stable sharing links

Claude Code style permission modes (default, acceptEdits, plan, bypassPermissions, auto) and user-defined permission rules for the pi coding agent. A more faithful port of Claude Code's harness than pi-permission-modes.

Temporary package placeholder for staged publishing

Local token and cost dashboard for pi sessions

Neon AI Gateway model provider for the pi coding agent

Write pi subagents and compose workflows, simply

Track and safely route Claude Code, Codex, Gemini, and Kiro subscription quotas across multiple accounts

OtterMind SQLX tools for the Pi coding agent: list datasources, test connections, execute SQL and open local result pages.

Pi extension that routes shell commands through RTK for token savings.

Pi extension for LunaRoute: login, model sync, request attribution, and first-class web tools.

Pi quality-of-life extension: compact statusline/π prompt, reliable multiline input, styled pasted-image chips, session naming/search/context import, scheduled prompts, handoff, permission prompts, notifications, custom compaction, and a collapsed-thinkin

A modular Pi extension for terminal UI, session workflows, adversarial review, and delegated agents

Universal LLM API client for 174 providers — chat, streaming, tools, embeddings, search, OCR, plus an OpenAI-compatible proxy and an MCP server.

Pi agent extension for Agent Plan — structured project planning in .planner/

Responsive Maestro operations sidebar, fallback status widgets, and footer for Pi.

Compact, reason-first tool output for pi with configurable layouts, live status, and diffs on demand.

Route closed judgements (pick one, rate this, yes/no) to a decision model instead of reading everything yourself. Typed, calibrated answers in ~300 ms for a fraction of a cent, via the `decide` CLI. Skills only, for Pi.

Pi permission extension — four permission modes (yolo/auto/approve/strict) with three-layer pipeline (AST + rules + AI classifier).

A pi coding-agent extension that connects a pi session to the local hail daemon so a paired phone can watch it and drive it.

Delegate separable work to autonomous subagents that run beside a PI Agent session and report back with evidence.

Pi extension: review the language you write to Pi while the agent does the real work, without ever blocking or rewriting the prompt.

Unified pi extension: hash-anchored read/edit/grep, structural code maps, AST-grep, file exploration (ls/find), and bash output compression