What can an AI agent do with Replicate?

A Replicate AI agent can submit predictions to any hosted model, track inference status, retrieve outputs, cancel stuck jobs, browse available models, and monitor your deployment fleet, all automatically on a schedule or triggered by events in other tools. Write actions like running models require your approval.

Which models can my agent run on Replicate?

Your agent can run any public model on Replicate, including image generators like Stable Diffusion and FLUX, video models, audio tools, and open-source language models. It can also run private models and fine-tunes hosted under your account.

How does my agent handle long-running predictions?

Replicate predictions can take seconds or minutes depending on the model. Your agent polls for completion, waits for results, and only proceeds once the output is ready, so you don't need to manage polling logic yourself.

Can I use Replicate alongside other tools in the same workflow?

Yes. A Replicate AI agent works best when combined with other integrations. For example, it can pull inputs from Google Drive, run a model on Replicate, and post the output to Slack, all in a single automated task.

Developer tools

Replicate, run by an agent.

An AI agent that runs, manages, and monitors your AI model deployments on Replicate.

Connect Replicate Book a demo

What needs your approval

Replicate never changes anything without you.

Anything that creates, edits or deletes in Replicate waits for your approval. Approve a recurring task once and it runs unattended after that.

Connecting Replicate

About 5 minutes

Auth: API key
Setup: About 5 minutes

Where to get access

Copy your API token from the Replicate account settings.

replicate.com

Example tasks

Real work people hand to an agent that uses Replicate.

generate product images from text prompts and save outputs to cloud storage on a schedule
run a fine-tuned language model over incoming support tickets and post summaries to slack
monitor long-running video generation predictions and notify your team when results are ready
batch process user-uploaded images through an upscaling model and return enhanced versions