Exact-error field guide
Search the visible text before replacing a key, model, runtime, or training setup. Matching happens in your browser, and the decoder deliberately strips long identifiers from comparison.
The decoder normalizes request IDs and punctuation, then ranks the closest authored guides. Matching happens in this browser.
401 Unauthorizedcloud provider429 rate limit or quota exceededEither DecaCap’s short-window abuse control or the upstream provider’s account quota rejected the request.429 Too Many Requestslocal companionConnection refused at the loopback endpointNothing accepted the browser’s request at the configured device-local address.ERR_CONNECTION_REFUSEDworkbenchCurrent entitlement does not permit the boxThe server rejected execution or saving outside the user’s current verified access.Your current entitlement permits Boxes 1–0.import exportGLB export has no source fileGLB export is a pass-through for a real imported GLB, not a 3D-generation feature.No real .glb file has been imported; DecaCap will not fabricate one.cloud providerKimi completed but final answer text is missingKimi returned a completed response, but its first choice did not contain extractable final message text.Kimi completed the request but did not include final answer text. Check that the model identifier is available to this API key.local companionLocal Companion is not pairedThe browser does not have a working loopback endpoint and pairing token for the user’s companion.Local companion is not paired.local companionLocal model ran out of memoryThe model, context, or training process requested more VRAM, unified memory, or system RAM than the device could safely provide.CUDA out of memorytrainingMLX-LM is not installed or detectedGoldCap prepared data, but the user’s local environment cannot launch the configured MLX-LM command.ModuleNotFoundError: No module named 'mlx_lm'cloud providerModel not found or unavailableThe provider accepted the connection but the named model is absent, retired, misspelled, region-limited, or unavailable to the account.model_not_foundlocal companionOllama model is missingOllama is reachable but the requested model is not installed under that exact local name.model not found, try pulling it firstcloud providerProvider completed without text outputThe upstream request completed, but the adapter could not extract answer text from the returned stream or final message.The provider completed without a text output.trainingTraining process finished but no adapter was foundThe local command exited, but the expected adapter files were not created at the configured path.Training completed but no adapter artifact was found.import exportWorkflow import is not valid JSONThe selected file could not be parsed as a DecaCap session or configuration.The workflow request is not valid JSON.14 authored diagnoses
The upstream request completed, but the adapter could not extract answer text from the returned stream or final message.
The provider completed without a text output.Kimi returned a completed response, but its first choice did not contain extractable final message text.
Kimi completed the request but did not include final answer text. Check that the model identifier is available to this API key.The provider rejected authentication before model work began.
401 UnauthorizedEither DecaCap’s short-window abuse control or the upstream provider’s account quota rejected the request.
429 Too Many RequestsThe provider accepted the connection but the named model is absent, retired, misspelled, region-limited, or unavailable to the account.
model_not_foundThe browser does not have a working loopback endpoint and pairing token for the user’s companion.
Local companion is not paired.Nothing accepted the browser’s request at the configured device-local address.
ERR_CONNECTION_REFUSEDOllama is reachable but the requested model is not installed under that exact local name.
model not found, try pulling it firstThe model, context, or training process requested more VRAM, unified memory, or system RAM than the device could safely provide.
CUDA out of memoryGoldCap prepared data, but the user’s local environment cannot launch the configured MLX-LM command.
ModuleNotFoundError: No module named 'mlx_lm'The local command exited, but the expected adapter files were not created at the configured path.
Training completed but no adapter artifact was found.The selected file could not be parsed as a DecaCap session or configuration.
The workflow request is not valid JSON.The server rejected execution or saving outside the user’s current verified access.
Your current entitlement permits Boxes 1–0.GLB export is a pass-through for a real imported GLB, not a 3D-generation feature.
No real .glb file has been imported; DecaCap will not fabricate one.