codebuddy and codebot-ai
About codebuddy
olasunkanmi-SE/codebuddy
An Autonomous AI Software Engineer
Operates as a multi-agent system using LangGraph DeepAgents with seven specialized subagents (Code Analyzer, Debugger, Tester, etc.), each with role-specific tool access, plus self-healing execution that automatically analyzes failures and retries until success. Integrates with 10 AI providers (Claude, GPT, Llama, etc.) with automatic failover, uses Tree-Sitter WASM for AST parsing across 7 languages, and extends via Model Context Protocol gateway supporting unlimited custom tools. Built on event-driven architecture with SQLite+vector store for codebase indexing, safety guardrails enforcing execution limits (2,000 events, 400 tool calls, 10-minute runtime), and human-in-the-loop approval for destructive operations.
About codebot-ai
zanderone1980/codebot-ai
Zero-dependency autonomous AI coding agent. Works with any LLM — Ollama, Claude, GPT, Gemini, DeepSeek, Groq, Mistral, Grok. Policy engine, Docker sandbox, 1,168 tests.
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work