# Replicate integration

> Run open-source and commercial AI models via a single cloud API, with async predictions and fine-tuning.

- Authentication: apiKey

## Tools (37)

- **Cancel Prediction**: Cancel a prediction that is still running.
- **Cancel Training**: Cancel a training that is still running.
- **Create Deployment**: Create a new deployment that runs a specific model version on chosen hardware.
- **Create File**: Upload a file that can be referenced as input to models.
- **Create Model**: Create a new model owned by a user or organization.
- **Create Prediction**: Run a prediction against a specific model version.
- **Create Prediction Using a Deployment**: Run a prediction against a deployment's current model version.
- **Create Prediction Using an Official Model**: Run a prediction against an official model's latest version.
- **Create Training**: Fine-tune a model by starting a training run on a base model version.
- **Delete Deployment**: Delete a deployment. Only deployments scaled to zero instances can be deleted.
- **Delete File**: Delete an uploaded file by ID.
- **Delete Model**: Delete a private model that has no versions and no predictions.
- **Delete Model Version**: Delete a model version and its associated predictions/output.
- **Download File**: Download the contents of an uploaded file via a signed URL.
- **Get Account**: Get the account (user or organization) that the API token belongs to.
- **Get Collection**: Get the models in a single collection by its slug.
- **Get Default Webhook Secret**: Get the signing secret used to verify default webhook requests.
- **Get Deployment**: Get information about a single deployment by owner and name.
- **Get File**: Get metadata for a single uploaded file by ID.
- **Get Model**: Get details for a single model, including its latest version and input schema.
- **Get Model README**: Get a model's README as Markdown.
- **Get Model Version**: Get details for a specific model version, including its input/output schema.
- **Get Prediction**: Get the current state and output of a prediction (poll this for async results).
- **Get Training**: Get the current state and output of a training (poll this for async results).
- **List Collections**: Get a paginated list of collections of models.
- **List Deployments**: List the deployments associated with the current account.
- **List Files**: List the files owned by the account.
- **List Hardware**: List the hardware SKUs available for running models.
- **List Model Examples**: List example predictions that showcase a model.
- **List Model Versions**: List all versions of a model.
- **List Predictions**: Get a paginated list of predictions created by the account.
- **List Public Models**: Get a paginated list of public models.
- **List Trainings**: Get a paginated list of trainings created by the account.
- **Search**: Search public models, collections, and docs with a text query (beta).
- **Search Public Models**: Search public models with a plain-text query using the QUERY method.
- **Update Deployment**: Update a deployment's version, hardware, or autoscaling instance bounds.
- **Update Model**: Update a model's metadata (description, README, and reference URLs).

Connect Replicate in General Input: https://www.generalinput.com/apps/replicate