Summary

ResearchClaw is a local-first Research OS centralizing and automating AI research workflows. It targets researchers and power users with a unified platform for literature review, note-taking, experiment tracking, and paper writing, preventing fragmentation across tools and chat histories, offering extensible skills and chat app integration.

How It Works

This project is a persistent runtime and workspace layer, built on FastAPI with a web console. Its core is a multi-agent runtime managing per-agent workspaces and binding rules, alongside a state layer tracking projects, workflows, tasks, artifacts, claims, and evidence. ResearchClaw standardizes extensibility via SKILL.md and MCP, integrating with external tools and chat platforms, consolidating scattered research into a durable, local-first system.

Quick Start & Requirements

Requires Python 3.10+. Install: git clone, cd ResearchClaw, pip install -e .. Initialize: researchclaw init --defaults --accept-security. Configure models (researchclaw models config or add), start service (researchclaw app --host 127.0.0.1 --port 8088). Frontend build steps (npm install, npm run build) may be needed. Docs: website/public/docs/intro.en.md, Roadmap: ROADMAP.md.

Highlighted Details

• Persistent Research State: Manages projects, workflows, tasks, artifacts, notes, claims, evidence, reminders locally.
• Multi-Agent Runtime: Features per-agent workspaces, binding rules, APIs for agents, channels, automation.
• Extensible Skills & Channels: Supports SKILL.md, Skills Hub, MCP; integrates with Telegram, Discord, iMessage, etc.
• Comprehensive Tooling: Built-in tools for literature search, BibTeX, LaTeX, data analysis, file ops, browser interaction.
• Claim/Evidence Graph: Links papers, notes, experiments, artifacts for verifiable research narratives.
• Experiment Tracking: Manages execution, validation, results.

Maintenance & Community

Part of the Research-Equality ecosystem. Official documentation and roadmap links provided. Specifics on active maintenance, community channels, or contributors are not detailed in the README.

Licensing & Compatibility

Licensed under Apache 2.0, permissive for commercial use and integration into closed-source projects.

Limitations & Caveats

Currently Alpha status. Areas for improvement: evidence-matrix quality, claim-evidence validation, external execution adapters, submission/reproducibility packaging. Incomplete as a full autonomous research platform.