Every action below is available as an MCP tool and a verb in Cerebral OS — callable from any AI agent, Claude, Cursor, Windsurf, or your own runtime via the BYOA API. All executions are governed, audited, and dry-run safe.
Cancel Prediction
replicate:cancel_prediction
Cancel a running prediction to stop processing and avoid further charges.
Write
High risk
Cancel Training
replicate:cancel_training
Cancel a running training job to stop processing and avoid further charges.
Write
High risk
Create Prediction
replicate:create_prediction
Create a new prediction (run a model) with the specified inputs.
Write
Medium risk
Create Training
replicate:create_training
Create a new training job to fine-tune a model with your data.
Write
Medium risk
Get Collection
replicate:get_collection
Get details about a specific model collection and its models.
Read
Low risk
Get Model
replicate:get_model
Get details about a specific model including its versions and schema.
Read
Low risk
Get Prediction
replicate:get_prediction
Get the current status and results of a prediction.
Read
Low risk
Get Training
replicate:get_training
Get the current status and results of a training job.
Read
Low risk
List Collections
replicate:list_collections
List featured model collections on Replicate.
Read
Low risk
List Models
replicate:list_models
List available models on Replicate with optional filtering.
Read
Low risk
List Predictions
replicate:list_predictions
List your predictions with pagination support.
Read
Low risk
List Trainings
replicate:list_trainings
List your training jobs with pagination support.
Read
Low risk