xuzhougeng/wisp-science
Open-source, local-first desktop AI research workbench for scientific computing with Python/R, MCP bioinformatics tools, SSH/WSL/GPU runtimes, and OpenAI/Anthropic models.
What it solves
Scientific research often requires switching between complex computational environments (Python, R), specialized biological databases, and various LLM providers. Wisp Science centralizes these into a single, local-first desktop application or headless CLI, preventing the fragmentation of research workflows and data.
How it works
It functions as an intelligent workbench that connects to OpenAI-compatible or Anthropic models. It manages persistent Python and R environments (locally, via SSH, or in WSL) and uses the Model Context Protocol (MCP) to interface with approximately 80 bioinformatics and computational biology databases. The system uses a multi-tier context management system and supports multi-agent workflows through a delegation system.
Who it’s for
- Computational biologists and bioinformaticians needing integrated access to specialized databases and coding runtimes.
- AI researchers looking for a local-first, privacy-conscious environment to manage agentic workflows and scientific data.
- Data scientists who require persistent, project-scoped execution environments for Python and R.
Highlights
- Hybrid Execution: Supports local, WSL, and remote SSH compute environments.
- Scientific Tooling: Bundled MCP servers for ~80 bioinformatics databases.
- Multi-Agent Workflows: Supports manual, assisted, and automatic multi-agent delegation.
- Persistent Runtimes: Maintains stateful Python and R REPLs for continuous computation.
- Cross-Platform: Available as a desktop app (Tauri) or a headless CLI.
- Secure Sync: Encrypted project synchronization between devices via a self-hosted relay.