#5252nd of 89 in Agent platformsFor science, health and research, science, research and papers
evoscientist
Multi-agent research assistant with persistent memory, CLI, WebUI and chat channels
- Stars
- 5k
- License
- Apache-2.0
- Last commit
- Nov 2026
- Last release
- Nov 2026
- Language
- Python
2,187 on 22 Apr 20264,960 on 11 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
EvoScientist is a Python research agent built on LangChain's DeepAgents. It runs six sub-agents (plan, research, code, debug, analyze, write) through an intake-to-verify workflow, and distills memory each turn into a knowledge graph that persists across sessions. It is driven from a CLI/TUI, a WebUI, a Windows desktop app, or chat channels such as Telegram and Slack.
Who it is for: Researchers wanting an autonomous, memory-keeping research agent
Strengths
- Six specialised sub-agents with a defined plan, execute, evaluate, write and verify flow
- Memory is distilled each turn and AutoSkills proposes reusable skills for review
- One agent session reachable from CLI, WebUI, Telegram, Slack, Feishu and WeChat
- Supports MCP servers, GitHub-installable skills and cron-style scheduled tasks
Weaknesses
- Listed providers are hosted APIs; local model support is not documented
- Fast release cadence with behavior changes, e.g. v0.4.0 project .env scope change
- macOS support limited to Apple silicon since v0.4.0
- Desktop app installer is listed only for Windows
What it needs
- no GPU
- Docker + Compose
- Needs LLM provider API key (Anthropic, OpenAI, Google, MiniMax, NVIDIA, OpenRouter and others), langgraph dev server, deepagents
- Models: Anthropic, OpenAI, Google, MiniMax, NVIDIA
What changed
“Breaking: a project .env holds only EvoSci command settings — provider keys, base URLs and other EVOSCIENTIST settings in a project .env are no longer read; it keeps only the EvoSci command's own set…”
CVE-2026-114603, GHSA-XCQ7-P3PF-P88H. “opentelemetry 1.45.2 — fixes CVE-2026-114603 (GHSA-xcq7-p3pf-p88h, unbounded tracestate parsing in opentelemetry-api).”
From the release notes. No flag means the note does not say so, not that the upgrade is safe.
Also in Agent platforms
See all 89| Rank | Project | Score |
|---|---|---|
| 1 | AutoGPTRole-based AI agents coordinated by one lead agent, plus a visual workflow builder | 95 out of 100 |
| 2 | daily_stock_analysisLLM-driven daily analysis of watchlist stocks across six markets, pushed to chat apps | 95 out of 100 |
| 3 | LangflowVisual builder for AI agents and workflows, exposed as API or MCP server | 94 out of 100 |
| #4 | deer-flowAgent harness running sub-agents, memory and sandboxes with pluggable skills | 93 out of 100 |
| #5 | DifyVisual builder for LLM workflows, RAG pipelines and agents, with APIs | 92 out of 100 |