Apps › AI

Replicate automation with Bizomate

The model for one particular job — an image model, a speech model for your language — is often one somebody published on Replicate. Bizomate runs it for each item: choose the model, fill in its own inputs as rows, and the result, files included, goes to the next step. It runs on your company's own Replicate account, and every run's record shows how long the model ran.

Start free

On every plan, Free included.

At a glance

Label	Value
Plans	Every plan
Connect	A Replicate API token
Starts runs	No trigger — starts from another app, a schedule or a webhook
Kind	Connected app · AI

What a workflow can do with it

Run a model
Runs any model on Replicate with its own inputs, and passes on what it makes

Choose from your own models and Replicate's official ones, or type any owner/name, such as black-forest-labs/flux-schnell. Leave Version empty for the latest, or pin one. The Inputs are read from the model's own description, required ones first; a file from an earlier step is sent to Replicate, and files the model makes come back as files. The step waits up to 60 seconds; a run still going then succeeds as processing, with its id.

Get a prediction
Reads a Replicate run that is still going, or one started elsewhere

By its Prediction ID — after a Wait, for a run that was still going.

How you connect it

Make an API token under Account settings › API tokens on Replicate and paste it into Bizomate under API token; the Replicate account needs billing set up before most models run. Bizomate tests the token before saving it by reading the account.

The token is your company's own, so Replicate bills your account for the time each model runs: Bizomate does not pay for or resell AI use, and each run's output shows how long the model ran. When a step runs, Bizomate sends Replicate what that step is given — the inputs, the files — under your account, and your account's terms with Replicate, and the model's own terms, apply to it.

The token is encrypted before it is saved, used only when a step runs or a test calls Replicate, and never shown again — not to you, your colleagues or Bizomate support. It belongs to one workspace, the Code step cannot read it, and revoking the connection wipes it at once.

In the help centre: Connect Replicate, Run a model on Replicate, Pause a run with Wait and Pass files between steps.

Questions about Replicate

Which models can I run?

Any model on Replicate. Choose from the list — your own models, then Replicate's official ones — or type a model's owner/name as Replicate shows it. Its inputs are read from the model's own description, so each model shows the inputs it takes.

What happens when a model takes longer than a minute?

Run a model waits up to 60 seconds. A run still going then succeeds with the status processing and its id; add a Wait step, then Get a prediction with that id to read the result.

Why pin a version?

A pinned version keeps giving the same kind of result when the model's owner changes the model. Leave Version empty to use the latest.

Who pays for the models?

Your company does: Replicate bills your own account for the time each model runs. Bizomate does not pay for or resell AI use. The model's own terms apply as well as your account's terms with Replicate.