Open Source AI Development Apps

The open-source API layer that turns local LLMs into production private AI applications with full Claude API compatibility

Build, test, and deploy multi-agent AI workflows with a visual editor, RAG data sources, MCP tool integration, and a production-ready REST API.

The state-of-the-art memory and context engine for AI agents — ranked #1 on all three major AI memory benchmarks.

Build AI-powered apps locally with your own keys, your own models, and zero cloud lock-in.

Turn coding agents into real teammates — assign issues, track progress, and compound reusable skills across a vendor-neutral, self-hosted platform.

Give your AI coding agents persistent memory — claude-mem captures everything that happens during sessions, compresses it with AI, and surfaces the right context in every future session automatically.

An incremental data indexing engine that keeps AI agent context perpetually fresh by reprocessing only what changed.

Open-source observability platform purpose-built for AI agents — trace, evaluate, debug, and monitor at scale with SQL access and real-time replay.

One API endpoint for 25+ LLM providers — route, track costs, enforce compliance, and switch models without changing your code.

Local-first AI memory with verbatim storage, pluggable backends, and 96.6% retrieval recall on LongMemEval — no API key required.

Orchestrate an army of AI coding agents—Claude Code, Codex, Gemini CLI, and more—running simultaneously in isolated git worktrees from a single Electron desktop app.

Run AI workloads at scale with a Pythonic serverless runtime that handles GPU inference, background jobs, and sandboxes with zero infrastructure overhead.

A native, Ghostty-based macOS terminal with vertical tabs, agent-aware notifications, and a scriptable browser built for running many parallel AI coding agent sessions instead of juggling tmux panes.

An open-source CLI that packages machine learning models into standard, production-ready Docker containers — no Dockerfile wrangling, no CUDA version hell.

Compress everything your AI agent reads — tool outputs, logs, RAG chunks, and files — before it reaches the LLM, achieving 60–95% fewer tokens with the same answers.