> ## Documentation Index > Fetch the complete documentation index at: https://docs.alakazam.gg/llms.txt > Use this file to discover all available pages before exploring further. # Read the current observation (passive) ## OpenAPI ````yaml /train-v1.yaml get /v1/sim/sessions/{id}/obs openapi: 3.1.0 info: title: Alakazam Train API version: 1.0.0 summary: >- Train policies inside a world model, certify them in a frozen physics oracle. description: > The **Train API** runs the dream-training pipeline as reproducible jobs: 1. TRAIN: CMA-ES evolves a 6-parameter controller *inside the doom world model* (the same public checkpoint as [`alakazamworld/doom-dungeon-hg`](https://huggingface.co/alakazamworld/doom-dungeon-hg) on Hugging Face, byte-identical, sha256-verified). 2. EXAM: the champion is replayed in a frozen Webots oracle (two physical worlds, fixed contract, anti-exploit control arms). The oracle is the sole scoreboard; dream fitness is never a capability claim. One `POST /jobs` does both and leaves a machine-readable `job.json` you can poll at any time, jobs are asynchronous, kill-tolerant, and append-only. ### Honest verdict semantics The exam verdict has four bars per world (contacts, clearance, coverage, path). Warm-started jobs reproduce the historical best-known profile, contact-safe (0 contacts in 40/40 episodes), but no policy in program history has passed all four bars. A `pass: false` verdict with clean contacts is the expected state of the art, not an error. ### Reproducibility Training is bit-deterministic for a given spec seed **on a fixed platform**; champions differ across platforms (macOS-arm64 vs Linux-amd64 float paths). Exam results are deterministic to aggregates (third-decimal physics drift). ### Availability & timings The serving VM runs on-demand (it is started for sessions, not 24/7). Measured wall-times, serving VM (e2-standard-2; CPU generation varies per boot): smoke job (`pop 6, gens 2, T 20`) 12-17 min; real-scale (`pop 10, gens 8, T 60`) 3-4 h; exam alone 4-6 min. Poll `GET /jobs/{job_id}`, never block on the POST. servers: - url: https://api.alakazam.gg/train description: > The Train jobs + exam API. Bearer-key auth; the serving VM behind it is start-on-demand, so coordinate a session window with us before calling. security: - bearerAuth: [] tags: - name: Jobs description: Submit and poll dream-training jobs. - name: Simulation gym description: > **Beta, self-hosted runner.** Drive a world headlessly as a training environment (gym) and stream the SNN-controller observation contract: two virtual proximity sensors, a terminal collision spike, object labels, and an optional square RGB camera. The same schema is emitted offline by the dataset exporter (JSONL episodes), so controllers trained on recorded data plug straight into the live loop. Each session holds a real world-model GPU stream, sessions are metered by wall-time; create only as many parallel environments as your session budget allows, and DELETE them when done. paths: /v1/sim/sessions/{id}/obs: servers: - url: '{runner_url}' description: > Simulation-gym endpoints are served by the self-hosted runner (gym-server.mjs), NOT by the jobs VM, your runner URL is provided with your key. variables: runner_url: default: https://your-runner-url.example get: tags: - Simulation gym summary: Read the current observation (passive) operationId: getSimObservation parameters: - name: id in: path required: true schema: type: string responses: '200': description: Current observation, no action applied. content: application/json: schema: $ref: '#/components/schemas/SimObservation' '404': description: Unknown session components: schemas: SimObservation: type: object description: > The SNN-controller observation contract. `proximity` = two virtual range sensors (0 = clear, 1.0 = contact range): compact obstacles grade by screen-space closeness, walls/elongated structures by time-to-contact; an object dead ahead excites both. `collision` is a TERMINAL SPIKE , true exactly once, on the step the contact engine confirms the hit , after which `done` stays true until reset. properties: proximity: type: object properties: left: type: number minimum: 0 maximum: 1 right: type: number minimum: 0 maximum: 1 collision: type: boolean description: > SPIKE, true exactly once per collision event (derived from the monotonic counter, so a pulse between two step reads is never missed). With terminal_collision=false the run continues and later events spike again. collision_count: type: integer description: Monotonic count of collision events this run. done: type: boolean description: >- Only latches when terminal_collision=true (or the world ends the run). t_ms: type: integer description: Milliseconds survived this run. sensor_age_ms: type: - integer - 'null' description: > Lag accounting, how stale the sensor side (proximity/labels) is relative to the camera frame in the same observation. Cloud-detector worlds run ≈500–1000 ms behind; local-detector worlds tens of ms. Controllers should compensate or discount accordingly. applied_action: type: string description: >- The discrete action actually applied (echoes wheel quantization). Present on step responses. labels: type: array items: type: object properties: id: type: string description: Persistent tracker id (`obj_`). type: type: string description: Normalized category (robot wall: null box: null barrier: null cone: null ...).: null confidence: type: number camera: type: object description: Present when the session requested a camera resolution. properties: image: type: string description: Square RGB frame as a data-URL JPEG. securitySchemes: bearerAuth: type: http scheme: bearer description: Single per-partner bearer key, provisioned by Alakazam. ````