A JavaScript client for Replicate. It lets you run models from Node.js and the web (coming soon!), and do various other things with Replicate.
This is an early alpha. The implementation might change between versions without warning. Please use at your own risk and pin to a specific version if you're relying on this for anything important!
Install with npm install replicate
Set your API token as an environment variable called REPLICATE_API_TOKEN
.
To run a prediction and return its output:
import replicate from "replicate";
const prediction = await replicate
.model(
"stability-ai/stable-diffusion:db21e45d3f7023abc2a46ee38a23973f6dce16bb082a930b0c49861f96d1e5bf"
)
.predict({
prompt: "painting of a cat by andy warhol",
});
console.log(prediction.output);
// "https://replicate.delivery/pbxt/oeJLu7D1Y7UWESpzerfINqgwZgONSCubSjSw0msf8i4AP2BCB/out-0.png"
If you want to do something like updating progress while the prediction is
running, you can pass in an onUpdate
callback function:
import replicate from "replicate";
await replicate
.model(
"stability-ai/stable-diffusion:db21e45d3f7023abc2a46ee38a23973f6dce16bb082a930b0c49861f96d1e5bf"
)
.predict(
{
prompt: "painting of a cat by andy warhol",
},
{
onUpdate: (prediction) => {
console.log(prediction.output);
},
}
);
If you'd prefer to control your own polling you can use the low-level
createPrediction()
method:
import replicate from "replicate";
const prediction = await replicate
.model(
"stability-ai/stable-diffusion:db21e45d3f7023abc2a46ee38a23973f6dce16bb082a930b0c49861f96d1e5bf"
)
.createPrediction({
prompt: "painting of a cat by andy warhol",
});
console.log(prediction.status); // "starting"
From there, you can fetch the current status of the prediction using
await prediction.load()
or await replicate.prediction(prediction.id).load()
.
While we'd love to accept contributions to this library, please open an issue before starting any new work so we can discuss the approach we'd like to take before you invest too much in writing code.