Replicate Swift client

This is a Swift client for Replicate. It lets you run models from your Swift code, and do various other things on Replicate.

To learn how to use it, take a look at our guide to building a SwiftUI app with Replicate.

Usage

Grab your API token from replicate.com/account and pass it to Client(token:):

import Foundation
import Replicate

let replicate = Replicate.Client(token: <#token#>)

[!WARNING] Don't store secrets in code or any other resources bundled with your app. Instead, fetch them from CloudKit or another server and store them in the keychain.

You can run a model and get its output:

let output = try await replicate.run(
    "stability-ai/stable-diffusion-3",
    ["prompt": "a 19th century portrait of a gentleman otter"]
) { prediction in
    // Print the prediction status after each update
    print(prediction.status)
}

print(output)
// ["https://replicate.delivery/yhqm/bh9SsjWXY3pGKJyQzYjQlsZPzcNZ4EYOeEsPjFytc5TjYeNTA/R8_SD3_00001_.webp"]

Or fetch a model by name and create a prediction against its latest version:

let model = try await replicate.getModel("stability-ai/stable-diffusion-3")
if let latestVersion = model.latestVersion {
    let prompt = "a 19th century portrait of a gentleman otter"
    let prediction = try await replicate.createPrediction(version: latestVersion.id,
                                                       input: ["prompt": "\(prompt)"],
                                                       wait: true)
    print(prediction.id)
    // "s654jhww3hrm60ch11v8t3zpkg"
    print(prediction.output)
    // ["https://replicate.delivery/yhqm/bh9SsjWXY3pGKJyQzYjQlsZPzcNZ4EYOeEsPjFytc5TjYeNTA/R8_SD3_00001_.webp"]
}

Some models, like tencentarc/gfpgan, receive images as inputs. To run a model that takes a file input you can pass either a URL to a publicly accessible file on the Internet or use the `uriEncoded(mimeType:) helper method to create a base64-encoded data URL from the contents of a local file.

let model = try await replicate.getModel("tencentarc/gfpgan")
if let latestVersion = model.latestVersion {
    let data = try! Data(contentsOf: URL(fileURLWithPath: "/path/to/image.jpg"))
    let mimeType = "image/jpeg"
    let prediction = try await replicate.createPrediction(version: latestVersion.id,
                                                       input: ["img": "\(data.uriEncoded(mimeType: mimeType))"])
    print(prediction.output)
    // ["https://replicate.delivery/mgxm/85f53415-0dc7-4703-891f-1e6f912119ad/output.png"]
}

You can start a model and run it in the background:

let model = replicate.getModel("kvfrans/clipdraw")

let prompt = "watercolor painting of an underwater submarine"
var prediction = replicate.createPrediction(version: model.latestVersion!.id,
                                         input: ["prompt": "\(prompt)"])
print(prediction.status)
// "starting"

try await prediction.wait(with: replicate)
print(prediction.status)
// "succeeded"

You can cancel a running prediction:

let model = replicate.getModel("kvfrans/clipdraw")

let prompt = "watercolor painting of an underwater submarine"
var prediction = replicate.createPrediction(version: model.latestVersion!.id,
                                            input: ["prompt": "\(prompt)"])
print(prediction.status)
// "starting"

try await prediction.cancel(with: replicate)
print(prediction.status)
// "canceled"

You can list all the predictions you've run:

var predictions: [Prediction] = []
var cursor: Replicate.Client.Pagination<Prediction>.Cursor?
let limit = 100

repeat {
    let page = try await replicate.getPredictions(cursor: cursor)
    predictions.append(contentsOf: page.results)
    cursor = page.next
} while predictions.count < limit && cursor != nil

Adding Replicate as a Dependency

To use the Replicate library in a Swift project, add it to the dependencies for your package and your target:

let package = Package(
    // name, platforms, products, etc.
    dependencies: [
        // other dependencies
        .package(url: "https://github.com/replicate/replicate-swift", from: "0.23.0"),
    ],
    targets: [
        .target(name: "<target>", dependencies: [
            // other dependencies
            .product(name: "Replicate", package: "replicate-swift"),
        ]),
        // other targets
    ]
)