What it does
The server exposes the Aethis decision engine, which compiles legislation, policy, contracts and regulation into deterministic logic, so the same input returns the same answer with an audit trail. Its 36 tools fall into six groups. Decision tools, including evaluation, schema, explanation and graph, are public and need no key. Discovery of your own tenant, rulebook management and rule authoring are in private beta. Decision calls can return a per-criterion evaluation trace, human-readable rule explanations, or a graph showing which criteria are satisfied for given inputs.
It suits checking eligibility against a published ruleset in a single decision call, or in a conversational flow where the tool returns the most informative remaining question after each answer until a decision is reachable. Teams drafting their own rulesets can use the authoring tools for generation, test refinement and publishing, and a separate skills package adds reusable workflows that call these tools.
Setup uses the aethis-cli command, which wires the server into Claude Code, Cursor, Claude Desktop or Windsurf and should be re-run after a key is rotated; the editor then needs a restart. Manual configuration is also documented for Cursor, Claude Desktop and Windsurf. For authoring access, a saved login profile is referenced through the AETHIS_PROFILE and XDG_CONFIG_HOME settings, so the host configuration does not hold the API key.
Limitations and requirements
Authoring and tenant tools require an invitation to the private beta. Rulebook decisions require an API key, and anonymous calls to them receive HTTP 401. Composition of rulesets through outcome logic is not covered by the current tools. The persona, conversational style and section transition hints are accepted but not yet acted on. XDG_CONFIG_HOME must be an absolute path, and relative values or credential symlinks that escape the home or config directory are not permitted.
Safety
Our team has not reviewed this resource. Only automated rules checked it; resources with high-risk signals are not published. These are signals, not a guarantee that installing it is safe. Try it in a sandbox project first.
- Runs code or changes an assistant's behaviour
Sources
- Official MCP Registry first seen 13 Nov 2026 · last seen 13 Nov 2026 · its page on the source
io.github.Aethis-ai/aethis-mcp
Deterministic eligibility decisions and test-driven rule authoring via the Aethis developer API.
v0.24.0As shown on registry.modelcontextprotocol.io when we last read it