MCP server

AskEngine ships a remote MCP server at https://askengine.dev/api/mcp — the same belief layer as the API, packaged as tools your coding agent discovers on its own. Connect once and you can ask things like "how are my studies doing?", "what do we currently believe about onboarding?", or "pull the evidence behind that theme" from inside your agent, and compose it with your other MCP servers ("check PostHog for this week's drop-off, then check AskEngine for what users said about it").

It authenticates with the same team API key as the REST API, and exposes the full v1 surface as fourteen tools. Ten reads — list_studies, get_study_readout, list_study_source_items, list_signals, get_signal, get_evidence, search_evidence, list_recommendations, get_recommendation, list_research_gaps — and the four recommendation lifecycle actions: decide_recommendation, record_recommendation_outcome, reopen_recommendation, and link_recommendation_work_item. Every tool returns a compact summary plus the raw API packet, so agents can answer quickly or drill down. The write tools are annotated as such, so clients like Claude Code ask before running them — and decisions are reversible via reopen_recommendation.

Connect

claude mcp add --transport http askengine https://askengine.dev/api/mcp \  --header "Authorization: Bearer $ASKENGINE_API_KEY"

Any other client that speaks Streamable HTTP works the same way: point it at the endpoint and send your API key as a bearer Authorization header.