BioClaw provides an AI-powered conversational interface for bioinformatics research, accessible via messaging platforms like WhatsApp. It addresses the fragmentation of research tools and literature by allowing researchers to perform complex analyses, data visualizations, and literature searches using natural language commands within group chats. This significantly streamlines workflows by eliminating the need to switch between multiple command-line tools, software, and databases.

How It Works

BioClaw is built upon the NanoClaw container-based agent architecture, integrating bioinformatics tools and domain knowledge from the STELLA project, and powered by the Claude Agent SDK. Core to its design is container isolation: each conversation group operates within its own Docker container, pre-equipped with a comprehensive suite of bioinformatics tools. Communication between the agent and orchestrator, as well as result delivery (images, plots, reports), occurs via filesystem inter-process communication (IPC). Per-group state, including messages and workspaces, is managed by a SQLite database, ensuring channel agnosticism through self-registration at startup.

Quick Start & Requirements

• Primary install / run command: A one-command setup script (scripts/setup.sh for macOS/Linux, scripts/setup.ps1 for Windows) is recommended.
• Non-default prerequisites: macOS / Linux / Windows (PowerShell 5.1+), Node.js 20+, Docker Desktop, and an Anthropic API key or OpenRouter API key are required.
• Links: NanoClaw, STELLA, Claude Agent SDK, docs/CHANNELS.md, docs/DASHBOARD.md.

Highlighted Details

• Messaging Channels: Supports WhatsApp (default), Feishu (Lark), WeCom, Discord, Slack (Socket Mode), WeChat Personal (experimental), and an optional local web UI.
• Bioinformatics Toolkit: Includes tools for sequence analysis (BLAST+, BWA, minimap2, BCFtools), quality control (FastQC, fastp, MultiQC), structural biology (PyMOL), data visualization (matplotlib, seaborn), literature search (PubMed integration), and more.
• AI Integration: Leverages Anthropic's Claude Agent SDK, with support for Anthropic or OpenAI-compatible model providers like OpenRouter.
• Containerized Workflows: Each chat group runs in an isolated Docker container, ensuring dependency management and reproducibility.

Maintenance & Community

The project provides a QR code for joining a WeChat group for discussion and exchange. Specific details on core maintainers, sponsorships, or a public roadmap are not detailed in the README.

Licensing & Compatibility

• License type: MIT License.
• Compatibility notes: The MIT license is permissive and generally compatible with commercial use and linking in closed-source projects.

Limitations & Caveats

Session history is preserved only within an active container session; after an idle timeout, a new container is initiated with a fresh context, potentially leading to a loss of conversational state.