Hokusai Vapor
Hokusai Vapor connects Hokusai 2 pipelines to Vapor 4. Decode an image upload, compose a recipe, and return a streamed image or PDF response.
Install
Section titled “Install”dependencies: [ .package(url: "https://github.com/vapor/vapor.git", from: "4.89.0"), .package(url: "https://github.com/ivantokar/hokusai-vapor.git", from: "2.0.0")]Add HokusaiVapor to the target that defines your routes. Hokusai’s native requirements remain the same: libvips 8.16+, Cairo with PDF support, and pkg-config on the host.
Configure once
Section titled “Configure once”import Vaporimport HokusaiVapor
func configure(_ app: Application) throws { try app.hokusai.configure() try routes(app)}configure() accepts an optional RuntimeConfiguration that freezes process-level settings before the first image, for example .init(maxConcurrentEvaluations: .fixed(4), maxQueuedEvaluations: 64). Without one, the defaults apply and the native runtime starts on first use. app.hokusai.capabilities() and app.hokusai.runtimeStatus report the runtime. Do not add a Vapor lifecycle shutdown hook.
First route
Section titled “First route”app.post("resize") { req async throws -> Response in try await req.hokusaiImage() .resize(width: 800, height: 600, fit: .cover) .jpeg(quality: 85) .response()}curl -X POST http://localhost:8080/resize \ --data-binary "@photo.jpg" \ -o output.jpgContinue with uploads and responses for raw bodies, multipart files, PDF output, and optional example routes.