> ## Documentation Index
> Fetch the complete documentation index at: https://docs.reactor.inc/llms.txt
> Use this file to discover all available pages before exploring further.

> ## Agent Instructions
> To build and serve your own model, start at /deploy/development/quickstart and /deploy/development/overview. Deploying is the default path: reactor init scaffolds a workspace, reactor auth login authenticates, and reactor model deploy registers the model, publishes the release with the weights/ folder, and activates it on Reactor's GPUs, in one command from that workspace. Docker must be running, because the publish step builds the image locally. Bump model.version in reactor.yaml before redeploying a change, because a release that already has an image is reactivated as it is. Deployment access is granted per account, so contact team@reactor.inc if a deploy is refused. Every key in reactor.yaml is documented at /deploy/platform/reactor-yaml. Model code imports reactor_runtime; Python client code imports reactor_sdk. The runtime overview explains the model interface. Running the model on your own machine with reactor run is optional and needs a GPU you attach with --gpus; /deploy/development/local-testing covers that loop and pairs a complete brightness model with a Python client test in a separate brightness-test workspace.
> Reactor hosts multiple models, each with its own connect slug (modelName) and command/event schema. The video model catalog — slug, typed SDK package, and links to its schema — is at /model-api-reference/overview. Robotics policy documentation starts at /robotics/overview; X-WAM observations, actions, and client integration are under /robotics/xwam/; Cosmos3 Nano Policy DROID is under /robotics/cosmos/nano-policy-droid/. Some models expose one slug per experience (e.g. HappyOyster); always take the slug from the model's own pages, never guess it.
> Fastest path to a working app: `npx create-reactor-app my-app --model=<slug>` scaffolds a complete app with secure auth wired up. Typed TypeScript SDKs are published as @reactor-models/<model>; Python uses the base reactor-sdk package.
> Auth: exchange an API key (rk_...) for a JWT via POST https://api.reactor.inc/tokens from your server. Never put the API key in client-side code.
> Append .md to any docs URL for clean Markdown. Search these docs via the MCP server at https://docs.reactor.inc/mcp.

# Troubleshooting

> Diagnose XR-1 RoboCasa365 history, action mapping, and execution-gate failures.

| Symptom | Likely cause | Correction |
| - | - | - |
| No first reply | Missing initial echo or fewer than four complete camera sets | Send `{"step": 0}` after valid state; publish all three views |
| First reply only | Execution count did not increase | Update observations, then send a larger integer echo |
| Echo advances but no reply | Missing video receipts or invalid state | Check four complete sets per accepted echo and four finite equal-width state rows |
| Valid JSON but poor actions | Incorrect 14-value layout | Use EE pose, two gripper positions, and base pose; schema prose mislabels this as two arms |
| History appears stale or shifted | Video repeated at wall-clock cadence or one track dropped frames | Compare capture indices across tracks and state; replace the synthetic publisher with an aligned adapter |
| Temporal spacing is wrong | Client and server both sampled history | Publish consecutive observations for default server sampling; verify actual selected indices |
| Simulator rejects actions | Full 60-wide row passed to controller | Take the first 12 values and use RoboCasa `convert_action` |
| Mode or base behavior is wrong | Incorrect action-column mapping | Use RoboCasa's base/torso/mode mapping and matching controller configuration |
| Old episode influences new one | Queued media or stored input survived reset | Clear client queues and reconnect with the new task and history |

Malformed state/history JSON is ignored by the model's readiness gate. An accepted command does not
prove that its values form a valid observation. Validate shape and finiteness before sending and
respect the [string-length limits](/robotics/xr1/robocasa365/reference#commands).

For diagnosis, log the release, prediction counter, executed-action count, camera capture indices,
sampled state indices, and command/reply timestamps. The server pairs arrival order, not capture
metadata. A network connection with no errors can still carry temporally misaligned observations.

For a minimal support case, start with the
[synthetic example](/robotics/xr1/robocasa365/quickstart). If it passes while a rollout fails,
inspect the observation/history adapter before changing the policy or its normalization.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.