API overview
One HTTP endpoint — run a decision, get a verdict slip back.
solvemax exposes one core endpoint: POST /api/v1/verdict. Send it a decision, get back a full
verdict slip — the exact same engine that runs the web UI, no scaled-down "API tier." The
proprietary pipeline (prompts, scoring, weighting) never leaves the server; you only ever see
this HTTP surface.
Base URL
Your deployment's own domain, e.g. https://solvemax.solveright.ai.
Endpoints
| Method | Path | Purpose |
|---|---|---|
POST | /api/v1/verdict | Run a full decision and get the verdict slip synchronously. |
GET | /api/v1/verdict?slug=... | Retrieve one of your past verdicts by its slug. |
See Authentication, The verdict endpoint, Rate limits, and Errors — or jump straight to the generated reference, built directly from solvemax's OpenAPI spec so it can never drift from what the endpoint actually accepts.
What's different from the web UI
- Synchronous, not streamed. The web UI streams live phase updates over Server-Sent Events; the API call blocks until the run finishes (or fails) and returns the complete JSON result in one response.
autoclarify by default. Instead of a round-trip for clarifying questions, the API defaults to the engine answering its own clarifying questions with its recommended picks — see Decision runs.- Always authenticated, always billed to your account credits — there's no anonymous-tier equivalent on the API.
Quick example
curl -X POST https://solvemax.solveright.ai/api/v1/verdict \
-H "Authorization: Bearer smx_live_..." \
-H "Content-Type: application/json" \
-d '{"decision": "Should I hire a senior generalist or two juniors with my next $150k of budget?"}'See The verdict endpoint for the full request/response shape.
Prefer not to write HTTP calls yourself?
The MCP server wraps this same endpoint as a tool any MCP-compatible AI assistant (Claude Desktop, Claude Code, Cursor) can call directly.