* [Claw Orchestrator v7.11.0](https://github.com/Enderfga/claw-orchestrator) – TypeScript runtime that runs Claude Code, Codex, Gemini, Cursor Agent and custom coding CLIs as persistent programmable headless agents.

* [PI-Desktop v0.18.0](https://github.com/vastsa/PI-Desktop) – Local-first desktop app for AI coding agents that runs plan, goal, and agent workflows while keeping models, keys, and conversation data on-device with user approvals.

* [Claw Orchestrator v7.10.0](https://github.com/Enderfga/claw-orchestrator) – TypeScript runtime that runs Claude Code, Codex, Gemini, Cursor Agent and custom coding CLIs as persistent programmable headless agents.

* [dsh-TUI v0.14.0](https://github.com/ccch1mneyyy/dsh-TUI) – Full-screen TUI plugin for DeepSeek Harness, offering Claude Code-like interactions, streaming thought expansion, agent task progress, and Esc-based rewind with TPS/token display.

* [Claw Orchestrator v7.9.0](https://github.com/Enderfga/claw-orchestrator) – TypeScript runtime that runs Claude Code, Codex, Gemini, Cursor Agent and custom coding CLIs as persistent programmable headless agents.

* [Claw Orchestrator v7.8.0](https://github.com/Enderfga/claw-orchestrator) – TypeScript runtime that runs Claude Code, Codex, Gemini, Cursor Agent and custom coding CLIs as persistent programmable headless agents.

* [PI-Desktop v0.17.0](https://github.com/vastsa/PI-Desktop) – Local-first desktop app for AI coding agents that runs plan, goal, and agent workflows while keeping models, keys, and conversation data on-device with user approvals.

* [PiClaw v3.3.0](https://github.com/rcarmo/piclaw) – Web-first self-hosted Debian sandbox packaging the Pi Coding Agent with streaming chat UI, persistent sessions, workspace tooling, and optional WhatsApp.

* [Claw Orchestrator v7.7.0](https://github.com/Enderfga/claw-orchestrator) – TypeScript runtime that runs Claude Code, Codex, Gemini, Cursor Agent and custom coding CLIs as persistent programmable headless agents.

* [Pi-Harness v1.9.0](https://github.com/wangmiaozero/pi-harness) – Desktop control center that visualizes and manages Pi Coding Agent Harness runtime state, sessions, models, skills, and workspace files with Git support.

* [Pi Agent Desktop v0.4.0](https://github.com/DLYZZT/pi-desktop) – Local-first Electron desktop workspace for managing Pi Coding Agent sessions, project files, browser tools, skills and plugins, with messaging integrations via Telegram and Lark.

* [Kangentic v0.45.0](https://github.com/Kangentic/kangentic) – Kanban board for coordinating AI coding agents, spawning and resuming sessions from one board with live status, worktrees, persistence, routing, and analytics.

* [pi-dynamic-workflows v3.14.0](https://github.com/QuintinShaw/pi-dynamic-workflows) – Claude Code–style dynamic workflows for Pi that orchestrate parallel routed subagents with verification, journaling resume, git worktree isolation, cost tracking, and a live TUI.

* [Kangentic v0.44.0](https://github.com/Kangentic/kangentic) – Kanban board for coordinating AI coding agents, spawning and resuming sessions from one board with live status, worktrees, persistence, routing, and analytics.

* [Late v2.0.1](https://github.com/mlhher/late-cli) – Deterministic coding agent orchestrator that spawns ephemeral subagents to perform exact-match code edits and avoid context bloat.

* [DevCodex v1.20.4](https://github.com/devcodex-labs/devcodex) – Intent-driven AI coding workflow runtime that keeps consistent context, skills, approvals, validation, and handoffs across six AI coding hosts.

* [PiClaw v3.3.2](https://github.com/rcarmo/piclaw) – Web-first self-hosted Debian sandbox packaging the Pi Coding Agent with streaming chat UI, persistent sessions, workspace tooling, and optional WhatsApp.

* [Harness Remote v3.1.2](https://github.com/giuliastro/harness-remote) – Companion app for controlling coding-agent harness sessions from phone or desktop with support for OpenCode and Oh My Pi.

* [Claudexor v3.25.1](https://github.com/razzant/claudexor) – Local-first control plane that runs multiple AI coding agents via a typed interface, adding quota-aware routing, best-of-N races, and auditable execution evidence.

* [PiClaw v3.3.1](https://github.com/rcarmo/piclaw) – Web-first self-hosted Debian sandbox packaging the Pi Coding Agent with streaming chat UI, persistent sessions, workspace tooling, and optional WhatsApp.

* [tmux-ide v2.9.5](https://github.com/wavyrai/tmux-ide) – tmux-ide adds an agent-aware terminal UI on top of existing tmux sessions, including live agent status, notifications, and crash-proof restore across clients.

* [Glowbom OSS v4.1.1](https://github.com/glowbom/glowbom-oss) – Coding-agent workflow that helps turn real software projects and prototypes into production-ready code using locally configured OpenCode providers and models.

* [tmux-ide v2.9.4](https://github.com/wavyrai/tmux-ide) – tmux-ide adds an agent-aware terminal UI on top of existing tmux sessions, including live agent status, notifications, and crash-proof restore across clients.

* [tmux-ide v2.9.3](https://github.com/wavyrai/tmux-ide) – tmux-ide adds an agent-aware terminal UI on top of existing tmux sessions, including live agent status, notifications, and crash-proof restore across clients.

* [zot v0.4.19](https://github.com/patriceckhart/zot) – Lightweight Go-based coding agent harness providing a single static binary, multiple provider integrations, basic tools, and extension support.

* [Claw Orchestrator v7.6.2](https://github.com/Enderfga/claw-orchestrator) – TypeScript runtime that runs Claude Code, Codex, Gemini, Cursor Agent and custom coding CLIs as persistent programmable headless agents.

* [Claudexor v3.20.1](https://github.com/razzant/claudexor) – Local-first control plane that runs multiple AI coding agents via a typed interface, adding quota-aware routing, best-of-N races, and auditable execution evidence.

* [Lynn v0.87.1](https://github.com/MerkyorLynn/Lynn) – Lynn is a desktop and CLI AI agent with real-time voice, long-term memory, and support for local Qwen GGUF models via llama.cpp.

* [zot v0.4.28](https://github.com/patriceckhart/zot) – Lightweight Go-based coding agent harness providing a single static binary, multiple provider integrations, basic tools, and extension support.

* [Superset desktop-canary](https://github.com/superset-sh/superset) – Terminal for running and managing 10+ parallel CLI coding agents locally.

* [zot v0.4.27](https://github.com/patriceckhart/zot) – Lightweight Go-based coding agent harness providing a single static binary, multiple provider integrations, basic tools, and extension support.

* [DSCODE v0.7.39](https://github.com/qiz029/dscode) – DSCODE is a macOS terminal coding agent built on DeepSeek Harness, using a persistent shared session runtime, session bridging, agent-to-agent task handoff, and independent reviews.

* [zot v0.4.26](https://github.com/patriceckhart/zot) – Lightweight Go-based coding agent harness providing a single static binary, multiple provider integrations, basic tools, and extension support.

* [Mixdog v1.0.12](https://github.com/tribgames/mixdog) – Windows desktop coding agent designed to match Codex performance while using substantially fewer tokens.

* [Freestyle 0.9.4](https://github.com/freestyle-voice/freestyle) – Local-first dictation app that converts speech to text via a hotkey, with optional cloud providers, transcription cleanup, and contextual corrections.

* [DSCODE v0.7.36](https://github.com/qiz029/dscode) – DSCODE is a macOS terminal coding agent built on DeepSeek Harness, using a persistent shared session runtime, session bridging, agent-to-agent task handoff, and independent reviews.

* [phi v0.28.4](https://github.com/pulseaiclub/phi) – Minimal Go terminal coding agent with sub-agents, hash-anchored file edits, and a permission gate plus on-demand MCP tool discovery.

* [San v1.24.1](https://github.com/genai-io/san) – Terminal-native unified runtime for specialized AI agents built on pluggable LLMs, search backends, personas, and skills.

* [omp GUI v0.9.18](https://github.com/nornzach/oh-my-pi-gui) – Electron desktop GUI that hosts parallel coding agents with split-session views, readable tool calls and diffs, model management, and usage analytics.

* [zot v0.4.22](https://github.com/patriceckhart/zot) – Lightweight Go-based coding agent harness providing a single static binary, multiple provider integrations, basic tools, and extension support.

* [DSCODE v0.7.35](https://github.com/qiz029/dscode) – DSCODE is a macOS terminal coding agent built on DeepSeek Harness, using a persistent shared session runtime, session bridging, agent-to-agent task handoff, and independent reviews.

* [DSCODE v0.7.34](https://github.com/qiz029/dscode) – DSCODE is a macOS terminal coding agent built on DeepSeek Harness, using a persistent shared session runtime, session bridging, agent-to-agent task handoff, and independent reviews.

* [DSCODE v0.7.33](https://github.com/qiz029/dscode) – DSCODE is a macOS terminal coding agent built on DeepSeek Harness, using a persistent shared session runtime, session bridging, agent-to-agent task handoff, and independent reviews.

* [Mixdog v1.0.11](https://github.com/tribgames/mixdog) – Windows desktop coding agent designed to match Codex performance while using substantially fewer tokens.

* [zot v0.4.21](https://github.com/patriceckhart/zot) – Lightweight Go-based coding agent harness providing a single static binary, multiple provider integrations, basic tools, and extension support.

* [Vicoa v0.1.38](https://github.com/vicoa-ai/vicoa) – Vicoa is an AI orchestrator that runs and coordinates multiple coding agents across devices, providing a unified session dashboard with parallel worktrees and remote control.

* [Open Vetta v0.5.61](https://github.com/openvetta/open-vetta) – Local-first desktop AI agent that brings models, tools, and project files into a controlled workspace with reviewable execution and extensible workflows.

* [omp GUI v0.9.17](https://github.com/nornzach/oh-my-pi-gui) – Electron desktop GUI that hosts parallel coding agents with split-session views, readable tool calls and diffs, model management, and usage analytics.

* [omp GUI v0.9.16](https://github.com/nornzach/oh-my-pi-gui) – Electron desktop GUI that hosts parallel coding agents with split-session views, readable tool calls and diffs, model management, and usage analytics.

* [phi v0.28.3](https://github.com/pulseaiclub/phi) – Minimal Go terminal coding agent with sub-agents, hash-anchored file edits, and a permission gate plus on-demand MCP tool discovery.

* [pi-task v0.42.56](https://github.com/mjasnikovs/pi-task) – Deterministic spec-orchestration for local models, running fixed refine→research→grill→compose→critique pipeline with crash-safe persisted checkpoints and bundled web/docs/fetch worker tools.

* [pi-task v0.42.55](https://github.com/mjasnikovs/pi-task) – Deterministic spec-orchestration for local models, running fixed refine→research→grill→compose→critique pipeline with crash-safe persisted checkpoints and bundled web/docs/fetch worker tools.

* [Mixdog effort-judge-v3](https://github.com/tribgames/mixdog) – Windows desktop coding agent designed to match Codex performance while using substantially fewer tokens.

* [pi-task v0.42.54](https://github.com/mjasnikovs/pi-task) – Deterministic spec-orchestration for local models, running fixed refine→research→grill→compose→critique pipeline with crash-safe persisted checkpoints and bundled web/docs/fetch worker tools.

* [pi-task v0.42.53](https://github.com/mjasnikovs/pi-task) – Deterministic spec-orchestration for local models, running fixed refine→research→grill→compose→critique pipeline with crash-safe persisted checkpoints and bundled web/docs/fetch worker tools.

* [Mixdog effort-judge-v2](https://github.com/tribgames/mixdog) – Windows desktop coding agent designed to match Codex performance while using substantially fewer tokens.

* [Mixdog effort-judge-v1](https://github.com/tribgames/mixdog) – Windows desktop coding agent designed to match Codex performance while using substantially fewer tokens.

* [AOCI-CODE v0.1.0-rc19](https://github.com/aoci-spec/aoci-code) – Indexing method that distills repository code and database knowledge into a portable, versioned cognition map for AI agents to read and act on consistently.

* [pi-task v0.42.49](https://github.com/mjasnikovs/pi-task) – Deterministic spec-orchestration for local models, running fixed refine→research→grill→compose→critique pipeline with crash-safe persisted checkpoints and bundled web/docs/fetch worker tools.

* [pi-task v0.42.48](https://github.com/mjasnikovs/pi-task) – Deterministic spec-orchestration for local models, running fixed refine→research→grill→compose→critique pipeline with crash-safe persisted checkpoints and bundled web/docs/fetch worker tools.

* [pi-task v0.42.47](https://github.com/mjasnikovs/pi-task) – Deterministic spec-orchestration for local models, running fixed refine→research→grill→compose→critique pipeline with crash-safe persisted checkpoints and bundled web/docs/fetch worker tools.

* [pi-task v0.42.46](https://github.com/mjasnikovs/pi-task) – Deterministic spec-orchestration for local models, running fixed refine→research→grill→compose→critique pipeline with crash-safe persisted checkpoints and bundled web/docs/fetch worker tools.

* [Radiant v0.9.66](https://github.com/templetongroup/radiant) – Mac local coding harness that supports agent chat across cloud and local models with live tool activity and an in-app terminal.

* [pi-task v0.42.45](https://github.com/mjasnikovs/pi-task) – Deterministic spec-orchestration for local models, running fixed refine→research→grill→compose→critique pipeline with crash-safe persisted checkpoints and bundled web/docs/fetch worker tools.

* [pi-task v0.42.44](https://github.com/mjasnikovs/pi-task) – Deterministic spec-orchestration for local models, running fixed refine→research→grill→compose→critique pipeline with crash-safe persisted checkpoints and bundled web/docs/fetch worker tools.

* [AOCI-CODE v0.1.0-rc18](https://github.com/aoci-spec/aoci-code) – Indexing method that distills repository code and database knowledge into a portable, versioned cognition map for AI agents to read and act on consistently.

* [pi-task v0.42.43](https://github.com/mjasnikovs/pi-task) – Deterministic spec-orchestration for local models, running fixed refine→research→grill→compose→critique pipeline with crash-safe persisted checkpoints and bundled web/docs/fetch worker tools.

* [pi-task v0.42.42](https://github.com/mjasnikovs/pi-task) – Deterministic spec-orchestration for local models, running fixed refine→research→grill→compose→critique pipeline with crash-safe persisted checkpoints and bundled web/docs/fetch worker tools.

* [senpi v2026.10.10](https://github.com/code-yeongyu/senpi) – Opinionated TypeScript monorepo fork of pi-mono that provides a curated coding-agent runtime with builtin extensions and core tweaks.

* [senpi v2026.10.9](https://github.com/code-yeongyu/senpi) – Opinionated TypeScript monorepo fork of pi-mono that provides a curated coding-agent runtime with builtin extensions and core tweaks.

* [Autoloom v0.1.0-beta.1](https://github.com/GanyuanRan/Autoloom) – Desktop AI coding client that embeds Aegis governance methods into the runtime, enforcing change-necessity checks, complexity review, causal diagnosis, and evidence-based delivery review.

* [senpi v2026.10.10-12](https://github.com/code-yeongyu/senpi) – Opinionated TypeScript monorepo fork of pi-mono that provides a curated coding-agent runtime with builtin extensions and core tweaks.

* [PI-Desktop v0.18.0-beta.5](https://github.com/vastsa/PI-Desktop) – Local-first desktop app for AI coding agents that runs plan, goal, and agent workflows while keeping models, keys, and conversation data on-device with user approvals.

* [PI-Desktop v0.18.0-beta.4](https://github.com/vastsa/PI-Desktop) – Local-first desktop app for AI coding agents that runs plan, goal, and agent workflows while keeping models, keys, and conversation data on-device with user approvals.

* [AOCI-CODE v0.1.0-rc20](https://github.com/aoci-spec/aoci-code) – Indexing method that distills repository code and database knowledge into a portable, versioned cognition map for AI agents to read and act on consistently.

* [CodeAF dev-20261009-973b7d2...](https://github.com/Agent-Field/CodeAF) – Coding harness that coordinates open-model agent work from a single terminal window with approval steps and cost controls across projects.

* [PI-Desktop v0.18.0-beta.3](https://github.com/vastsa/PI-Desktop) – Local-first desktop app for AI coding agents that runs plan, goal, and agent workflows while keeping models, keys, and conversation data on-device with user approvals.

* [CodeAF dev-20261008-8761488...](https://github.com/Agent-Field/CodeAF) – Coding harness that coordinates open-model agent work from a single terminal window with approval steps and cost controls across projects.

* [Autoloom v0.1.0-alpha.13](https://github.com/GanyuanRan/Autoloom) – Desktop AI coding client that embeds Aegis governance methods into the runtime, enforcing change-necessity checks, complexity review, causal diagnosis, and evidence-based delivery review.

* [wuu v2026.10.6](https://github.com/blueberrycongee/wuu) – Local-project AI agent desktop app that coordinates model providers to read code, apply changes, run commands, and review diffs and results.

* [Autoloom v0.1.0-alpha.12](https://github.com/GanyuanRan/Autoloom) – Desktop AI coding client that embeds Aegis governance methods into the runtime, enforcing change-necessity checks, complexity review, causal diagnosis, and evidence-based delivery review.