omp packages

Token rate status extension for pi-coding-agent - shows average output tokens per second

Keep Pi models fresh: /model-sync discovers live provider catalogs, enriches from models.dev, and writes missing models to models.json.

Model-agnostic, local-first persistent memory for the Pi coding agent — Fable 5's memory structure minus the overzealous parts

🛂 Use Claude models in pi through relays that only accept Claude Code traffic (e.g. anyrouter). In-process request shim, no proxy.

Lex — BMW Legal AI Agent Platform built on Pi

Interactive Pi extension to manage skill loading: per-skill toggles at global and per-project scope, with project-level three-state (disable/enable/inherit).

Vega-Lite chart extension for pi coding agent - render data visualizations as inline images

Automatically name and keep Pi sessions organized

Spec-driven development workflow extension for pi

Socratic planning and shared-understanding sessions for pi.

Local multi-provider AI token spend dashboard for Claude Code, Codex, OpenCode, Pi, OMP, and Cursor.

TinyFish for the Pi coding agent — free web search and page reading, plus goal-driven browser automation on live and logged-in sites. Skills work standalone; the bundled MCP server activates if you also install pi-mcp-adapter.

Dev-only visual annotation loop: point at any localhost preview, click an element, file a durable change request.

Pantheon-style multi-agent orchestration for pi, inspired by oh-my-opencode-slim

SKILL.state + Jev for Pi: execution state maintained by typed decisions and code, separate from the reasoning LLM.

Pi package that bundles the Object-Oriented Unified Process extension and skills for the pi coding agent.

Show the current sub2api site's balance and today's actual spend in the Pi footer (read-only)

Opinionated multi-provider coding harness for the pi agent

Host-neutral semantic codebase search with embeddings, symbol discovery, and call-graph tooling

Antigravity / Cloud Code Assist provider for Pi with multi-account quota failover

Build knowledge artifacts that survive revision and session boundaries — plans, specs, repo or data analyses, reports — with declarative seams, a checkpoint, provenance, and a domain-agnostic patch tool.

Shared utilities for pi-atelier extensions — settings, project config, memory parsing, tool output truncation, and more

Pi extension that strips global AGENTS.md and CLAUDE.md from the effective prompt for selected projects.

Custom /m model picker for pi with favorites pinned above a divider and hide support

Render images inline in the pi TUI (reads, browser screenshots) without spending model tokens. Strips pixels from tool results before they reach the model.

Slot-configurable Ghostty tab title for pi coding agent — braille spinner, LLM session summary, tool tracking, native OSC 9;4 progress bar

Zero-Token-Waste Memory Engine, Knowledge Graph & Notes System for AI Coding Agents via MCP.

Command Code Provider API for Pi — every top model (Claude, GPT, Gemini, DeepSeek, Kimi, GLM, Qwen, Grok & more) behind one OpenAI- or Anthropic-compatible endpoint, with the live model catalog discovered automatically at startup.

Asynchronous delegation from Pi to independent child Pi sessions.

GPT-5.6, Codex and Pi prompt-caching skill.

长篇网文创作系统 Pi 扩展：/webnovel-* 命令 + webnovel_run 工具 + 技能/提示词资源。

Pi extension that exposes VS Code editor context through Claude Code's local IDE bridge.

Shell dock and /shell inspector for pi's background bash jobs

Pi package that adds an interactive ask_user clarification tool.

Multi-backend subagents and dynamic workflow orchestration for pi.

Jev-driven context pruning for pi: replaces stale tool results with recoverable stubs before each LLM call

Pi extension. A structured questionnaire the model can put to you when it would otherwise guess, with typed options instead of free-form replies.

Pre-write syntax validation for pi using tree-sitter WASM grammars

OpenCode-style command aliases for Pi, such as /exit, /clear, and /models.

Show Pi lifecycle state in a tmux pane picker

A focused Vim-style editing layer for pi, with motions, text objects, visual selections, and clipboard sync

One-way attention notifications for Pi via terminal OSC 777, optional tmux bell, and configurable JSON webhooks.

Manually export saved Pi conversations to Langfuse, incrementally and without background tracing.

Kill & Handoff protocol for pi — checkpointed context resets, watermark-enforced landing, and fresh-session task chaining for multi-agent workflows

Remote observation and control panel for pi coding agent

Protect Pi agent shell commands and file mutations with configurable block and confirmation rules.

Client-side context layers for pi: microcompact placeholders and model-driven history snipping

Fast deep research for the pi coding agent: reads up to 100 pages in full per round, Jev keeps only the passages that answer your questions.

Multi-line status bar for the pi coding agent — Claude Code-style identity line + meters line, with dual TPS meters (persistent sparkline + live gauge)

Voluntary Pi compaction from 60% context with temporary branch-local session memory.