Pi Extensible Workflows is an extension for developers who want Pi to coordinate several AI agents through repeatable, resumable workflows. The orchestration runs in your Pi host environment using Node.js, with the same access to files and processes as Pi. Its workflow scripts define task order, parallel work, and approval checkpoints.
Workflows can combine agent tasks with shell checks, split reviews across agents, and isolate work in Git worktrees. Agents can retain their conversation across turns, so an ongoing task doesn't have to start with a fresh transcript each time. Reusable workflows let you keep a process for later tasks instead of asking the main agent to write it again.
The included reviewLoop pairs implementation with review. Packaged roles cover development, code review, scouting, and research, with customizable model choices, tools, skills, and prompts. Those roles can also launch ordinary Pi sessions without a workflow. Standalone subagents support inspection, steering, stopping, and retries.
You can inspect running work. The terminal interface shows the agent tree, costs, and workflow script, while Trajectory plots model and tool events on a timeline. The Herdr companion opens workflow agents in separate panes for closer inspection. A separate CLI supports headless runs, inspection, and exports.
Claim this page and we'll verify you by hand. Pi Extensible Workflows gets the verified badge, and you can upgrade the listing to be featured on localhosted. Proud to be listed? Put our badge on your site.
Want more people to find Pi Extensible Workflows?Promote it
Something wrong or outdated on this page?
2.3KUpdated 2 days agoApache-2.0
macOS · Windows · Linux · Docker · Web#Agent Skills#Guardrails#Human approval
Agency Orchestrator turns a task description into a team of specialist AI roles that divide the work and combine their results. It's for solo builders, developers and content creators who want repeatable workflows for product planning, research or code review. The open-source software uses the Apache 2.0 license and runs locally through a browser interface, a desktop app for macOS, Windows and Linux, or Docker on a server or NAS.
3.9KUpdated 1 day agoApache-2.0
#Agent Client Protocol#Agent Skills#Code execution
403Updated 12 hours agoApache-2.0
macOS · Windows · Linux#Agent Skills#Code execution#Git integration
409Updated 1 day agoApache-2.0
#Code execution#Guardrails#Human approval
556Updated 3 weeks agoApache-2.0
Web · VS Code#Human approval#LLM tracing#MCP
28.5KUpdated 8 hours ago
Web#Human approval#LLM tracing#MCP
Fast Agent is a terminal coding assistant and Python toolkit for developers who want to build and evaluate AI agents with their choice of model. It supports local LLMs through a generic provider and automatically configures connections to llama.cpp servers. The toolkit runs on your machine; choosing Anthropic, Google, or a remote OpenAI-compatible endpoint sends model requests to that service.
Liza coordinates coding agents through a software delivery workflow with separate authors and reviewers. It's for developers who want to delegate work from specifications through tested, documented code while retaining control over product decisions. Its Go supervisor enforces role boundaries, validation gates and merge authority in code, rather than relying entirely on agents to follow prompts.
MartinLoop is a free, self-hosted runtime for developers and engineering teams delegating software work to AI coding agents. It wraps Claude Code, Codex CLI and Gemini CLI with limits and checks that sit outside the agent. The agent writes the code; MartinLoop decides when work can continue and when the result passes the configured checks.
MASFactory is an open-source Python framework for developers building workflows in which multiple AI agents work together. Its Vibe Graphing feature drafts a collaboration graph from a plain-language description, then lets people refine and approve the design through dialogue before it becomes an executable workflow. The license is Apache 2.0.
Mastra is a TypeScript framework for developers building AI agents and applications on their own servers or inside existing web apps. Its server runs locally or as a standalone deployment; Mastra Cloud provides a hosted alternative. Model routing connects to providers such as OpenAI, Anthropic, and Gemini, so running the framework locally doesn't keep those model requests on your machine.