EvoScientist — Multi-Agent System for Autonomous Scientific Discovery
EvoScientist is an open-source multi-agent platform that enables self-evolving AI scientists to autonomously conduct end-to-end scientific research. Trending on HuggingFace with 545 upvotes, it coordinates three specialized agents: a Researcher Agent for idea generation, an Engineer Agent for experiment implementation, and an Evolution Manager Agent that distills insights from prior interactions into reusable knowledge.
The system deploys six sub-agents (planning, research, coding, debugging, analysis, writing) working in concert through a structured scientific workflow. It supports multiple LLM providers (Claude, GPT, Gemini, Nemotron) and integrates with MCP servers for extensibility. Multi-channel interfaces include CLI, Telegram, Slack, Discord, and WeChat.
Results speak for themselves: EvoScientist ranked #1 on DeepResearch Bench II at submission time, and all six full papers it generated were accepted at ICAIS 2025, with two receiving major awards (Best Paper and AI Reviewer's Appraisal Award).
Licensed under Apache 2.0 and available via PyPI (pip install EvoScientist), it represents a new paradigm where AI agents don't just assist researchers but evolve into autonomous research collaborators.
GitHub: https://github.com/EvoScientist/EvoScientist
Paper: https://arxiv.org/abs/2603.08127
← Back to all articles
The system deploys six sub-agents (planning, research, coding, debugging, analysis, writing) working in concert through a structured scientific workflow. It supports multiple LLM providers (Claude, GPT, Gemini, Nemotron) and integrates with MCP servers for extensibility. Multi-channel interfaces include CLI, Telegram, Slack, Discord, and WeChat.
Results speak for themselves: EvoScientist ranked #1 on DeepResearch Bench II at submission time, and all six full papers it generated were accepted at ICAIS 2025, with two receiving major awards (Best Paper and AI Reviewer's Appraisal Award).
Licensed under Apache 2.0 and available via PyPI (pip install EvoScientist), it represents a new paradigm where AI agents don't just assist researchers but evolve into autonomous research collaborators.
GitHub: https://github.com/EvoScientist/EvoScientist
Paper: https://arxiv.org/abs/2603.08127