EvoScientist is a self-evolving AI scientist that plans, researches, codes, analyzes, and writes with a multi-agent system. EvoScientist features a dynamic multi-agent system where the main agent composes synchronous and asynchronous workers on demand from specialized agents, general-purpose agents, Skills, and MCP tools, with no fixed agent count. EvoScientist includes persistent memory that survives across sessions, self-linking into a knowledge graph that marks relationships as complements, contradicts, or supersedes.
EvoScientist provides a 6-phase scientific workflow: Intake, Plan, Execute, Evaluate, Write, and Verify. EvoScientist supports MCP servers and installable research-lifecycle skills from GitHub, and is compatible with other AI coding agents out of the box.
No comments yet.