Open Source AI Agents
Discover 404 open source AI Agents repositories from GitHub, each with an analysis of what it does, key features, use cases and alternatives. AI Agents projects here are most often combined with Python, Developer Tools and MCP. Last updated October 4, 2026.
404 repositories · updated October 4, 2026
ralph-orchestrator: Coordinate AI Agents in Task-Completion Loops
Ralph Orchestrator runs AI coding agents in structured loops, using specialized roles and quality gates to keep work moving toward completion. It suits developers who want repeatable, multi-step workflows across several agent CLIs.
agentfs: Store and Audit AI Agent State in SQLite
AgentFS stores an agent’s files, key-value state, and tool-call history in a SQLite database. It is aimed at developers who need portable, queryable agent state and filesystem access through SDKs or a CLI.
cq: Share Collective Learning Between AI Agents
cq is a knowledge store and workflow for coding agents to query, contribute, and validate reusable lessons. It helps teams and individual developers avoid repeating debugging work across agent sessions and machines.
cli: Connect AI Agents to Models and APIs with MCP
SandBase CLI connects supported AI clients to a catalog of AI models and APIs through a local MCP bridge. It is aimed at developers who want one integration for discovery and execution instead of configuring individual services in each client.
tessera: Organize AI Coding Sessions Across Projects
Tessera is a local workspace for coordinating Claude Code, Codex, and OpenCode sessions. It brings parallel agent tasks, Git worktrees, session views, and Git workflows together for developers managing several coding tasks at once.
agent.cpp: Build Local AI Agents in C++
agent.cpp is a C++ library for building agents that run small language models locally through llama.cpp. It provides an agent loop, tools, callbacks, and support for GGUF models, grammar constraints, and LoRA adapters.
wasm-agents-blueprint: Run Python AI Agents in the Browser
Mozilla AI's Wasm Agents Blueprint runs the OpenAI Agents Python SDK in a browser using Pyodide and WebAssembly. It suits developers exploring browser-based agents with hosted or local LLMs, without deploying an agent server.
agent-factory: Generate Executable AI Agent Workflows
Agent Factory turns natural-language task descriptions into executable Python agent workflows, using MCP tools and the any-agent library. It is aimed at developers who want to create, evaluate, and run agents without hand-coding every workflow.
envharness: Adapt Environments for Agent Learning
EnvHarness wraps frozen interactive environments with composable layers that can target an AI agent’s weaknesses without changing benchmark code or grading. It is for researchers building and evaluating agent-training workflows across different environments.
agenttrail: Observe AI Coding Agents Locally
Agenttrail turns local file changes and supported coding-agent activity into project maps and a 3D task view. It helps developers follow progress and spot work that needs attention without managing agents or sending data to a cloud service.
magic-context: Manage Coding Agent Context and Memory
Magic Context manages coding-agent context and project memory across sessions, reducing reliance on disruptive compaction and manual note-taking. It is designed for developers using OpenCode, Pi, or Oh My Pi who want persistent, searchable project knowledge.
aft: Give Coding Agents IDE-Style Code Tools
AFT equips OpenCode, Pi, and OMP coding agents with symbol-aware code navigation and edits, semantic search, and persistent shell tasks. It is a fit for developers who want agents to work across a codebase more precisely and with less noisy tool output.