MCP server¶
Coming soon
A Model Context Protocol (MCP) server for TwinCell is on the roadmap. It will let MCP-compatible agents and assistants (e.g. in your IDE or chat client) drive TwinCell directly — running target validation (with its causal mechanism) and (later) simulation as tools, without writing client code.
What to expect¶
- Expose the core capabilities as MCP tools:
- Target validation — score candidate targets for efficacy against a state change.
- MoA elucidation — return intermediary nodes and causal paths between a target and DEGs.
- Simulation (coming later) — predict DEGs from a target and control state.
- Authenticate with your existing
DEEPLIFE_API_KEY. - Stream long-running prediction status back to the agent.
In the meantime¶
Use the Python toolkit directly — see the Quick Start and API Reference. Want early access or to shape the MCP design? Reach out via Access & Support.