Hesperan provides an MCP server that enables AI agents to make calibrated decisions by asking typed questions (choice, yes/no, score) and receiving probabilistic answers. Agents can act on high-confidence results or defer to users, supporting tools like decide, decide_with_profile, and report_outcome for structured judgment workflows.
- Operator: Hesperan
- Endpoint: https://api.hesperan.com/mcp
- Transport: streamable-http
- Sign-in: No sign-in
- Category: AI and machine learning
- Use cases: Ask agents to evaluate risky shell commands before execution; Get calibrated probabilities for decision profiles; Report outcomes to improve future decision accuracy
- Tools: decide, decide_with_profile, report_outcome
- Documentation: https://hesperan.com/docs/mcp
- Directory listing: https://koskamo.com/explore/hesperan-mcp
- Also known as: Hesperan MCP MCP; hesperan-mcp