SWE-agent
swe-agent.com
Open-source autonomous software engineering agent from Princeton. Uses a custom Agent-Computer Interface (ACI) optimized for LLM code editing tasks. Set state-of-the-art performance on SWE-bench — the standard benchmark for autonomous bug fixing in real GitHub repos.
More in Vertical Autonomous Applications
Self-hosted AI coding assistant with code completion, chat, and agent capabilities. Runs on your own hardware or private cloud with consumer GPU support. Provides the core features of GitHub Copilot without sending code to third parties.
Free AI code acceleration toolkit with code completion and search across 70+ languages and 40+ IDEs. Uses its own models to provide fast, accurate suggestions. Windsurf IDE from Codeium ships a full agent mode with multi-file editing.
Open-source system for generating and iteratively executing code, inspired by GPT-4 Code Interpreter. Fine-tuned models that excel at data analysis, visualization, and math. Provides a self-hosted alternative to commercial code interpreters.
Open-source AI code assistant for VS Code and JetBrains with full agent mode. Supports any LLM (local or cloud) and lets you define custom slash commands and context providers. Community-driven alternative to Cursor with complete model flexibility.
Building autonomous AI employees starting with Ava, an AI sales development representative. Ava researches prospects, writes personalized outreach emails, manages follow-up sequences, and books meetings — running entire outbound sales workflows without human intervention.
Open-source platform for AI software agents (formerly OpenDevin). Agents can write code, run commands, browse the web, and call APIs to complete full software engineering tasks. Ships with a web UI and supports 10+ agent strategies.