selfhost.directory

The index of self-hostable projects — ranked, categorized and served with live updates.
Own your data and stop giving it to big tech.

ScienceClaw logo

Self-host ScienceClaw

v0.0.4

Personal research assistant with 1,900+ scientific tools

BetaOpen SourceMIT
Updated7 months ago
LanguagePython
Setup difficultyModerate
Stars
675
RAM
8 GB
Ports
8000 (backend REST API), 8080 (sandbox API), 8068 (websearch API), 8001 (task scheduler API)
Requires
MongoDB, Redis, SearXNG
Docker
swr.cn-north-4.myhuaweicloud.com/claw/scienceclaw-backend
Install with Docker Compose
Login page with default admin credentials

What is ScienceClaw?

This personal research assistant combines LangChain DeepAgents with AIO Sandbox to deliver a local-first AI agent for scientific work. It ships with over 1,900 pre-integrated tools covering drug discovery, astronomy, earth science, chemistry, biodiversity, social science, academic literature, and data science. Users can generate polished reports in PDF, Word, PowerPoint, and Excel formats, create custom tools and skills through natural language, and automate recurring tasks with cron-style scheduling. A built-in file browser and resource dashboard provide visibility into workspace contents and system health. All code runs inside a Docker sandbox that cannot reach the host filesystem or environment, and data remains in a local workspace directory. Deployment options include a 10-service Docker Compose stack (frontend, backend, sandbox, MongoDB, Redis, SearXNG, websearch, task scheduler, Celery worker, Celery beat) or a Windows desktop installer. Default login is admin/admin123. The backend uses MongoDB for persistence, Redis for task queuing, and DeepSeek as the default LLM provider.

Should you self-host ScienceClaw?

ScienceClaw scores 50/100 on selfhost.directory (how scoring works) and ranks #40 of 120 in AI Assistant. It has 675 GitHub stars.

There was no stable release in the past 12 months, so check the repository's recent activity before you rely on it.

Setup is moderate: expect some configuration beyond a single command. An official Docker image is available. It is MIT-licensed and written mostly in Python.

Its closest alternative on selfhost.directory is hermes-agent (89/100, 257 885 stars). It scores higher overall, so compare the two before you pick. See ScienceClaw vs hermes-agent side by side.

Features

  • 1,900+ built-in scientific tools
  • Custom tools and skills system
  • Feishu/Lark Integration
  • File management system
  • Isolated sandbox execution
  • Multi-format report generation
  • Resource Monitoring
  • Scheduled Tasks
  • Web search and crawl

ScienceClaw README

How to install ScienceClaw

Install ScienceClaw with Docker Compose

Environment variables
API_KEY, API_BASE, MODEL_NAME, TEMPERATURE, MAX_TOKENS, LOG_LEVEL, LARK_APP_ID, LARK_APP_SECRET, TASK_SERVICE_API_KEY, DISPLAY_TIMEZONE

git clone https://github.com/AgentTeam-TaichuAI/ScienceClaw.git cd ScienceClaw docker compose -f docker-compose-release.yml up -d --pull always

Pulls pre-built images; no local compilation. Open http://localhost:5173

More in Data Analysis

Rate this project

Click to rate — no account needed.

0 comments

Posting as guest
Loading comments…