{"schema_version":"v1","name_for_human":"SeaOtter — OtterScore","name_for_model":"seaotter_otterscore","description_for_human":"Grade your agent's work against an acceptance policy before you ship it. OtterScore is a hostile-by-default critic: ship / route to fix / quarantine / block.","description_for_model":"SeaOtter / OtterScore grades agent work (code, text, documents, decks, spreadsheets, images, video) and its trajectory against an acceptance policy and returns a score (0.0-1.0, 1.0 = ship, 0.0 = must block; lower = more flawed), a band (ship / route_to_fix / quarantine / block), located flaws[], and concrete upgrades[]. Use it to check work before delivering it. Get a free key with no human via POST /api/v1/agent-keys/signup {email, org_name?}. The robust grading entry is the async jobs path (POST /api/v1/eval/jobs -> poll GET /api/v1/eval/jobs/{job_id}) which tolerates the scale-to-zero GPU cold-start (warm grades are seconds; a cold first grade can take several minutes (up to ~6 min) while the GPU loads the model); a minimal body {user_prompt, artifact_parts} works. Read https://seaotter.ai/llms.txt for the full agent contract.","auth":{"type":"user_http","authorization_type":"bearer","instructions":"Authenticate every /api/v1/eval/* call with Authorization: Bearer <sk-otter-...>. Mint a free key with POST /api/v1/agent-keys/signup or at https://seaotter.ai/developers."},"api":{"type":"openapi","url":"https://api.seaotter.ai/api/v1/openapi.json","is_user_authenticated":true},"logo_url":"https://seaotter.ai/brand/seaotter-otter.png","contact_email":"hello@seaotter.ai","legal_info_url":"https://seaotter.ai/terms","llms_txt":"https://seaotter.ai/llms.txt","agent_manifest":"https://seaotter.ai/.well-known/ai-agent.json"}