My SDK docs are code samples in someone else’s docs site. What then?

Octri writes and hosts the docs from the same spec, with the call through your SDK on every endpoint page. If those pages live on Mintlify or ReadMe today, their guides here cover moving the written pages too.

Does trying the import change my current setup?

No. The importer only reads the files you upload, and Pull from GitHub only reads your repository. Nothing is written back to either, so your current site keeps running until you move the domain.

Will someone help with the move?

Yes. Book a call and a founder will go through your project with you and help the first release over by hand. Questions also reach us at founders@octri.dev.

Migration guide

Migrate from Speakeasy to Octri

Upload your Speakeasy workflow with the spec it reads. Your languages, SDK name and package name come across, with method names, pagination and streaming from the x-speakeasy extensions already in your spec.

Start on the Free plan Book a call

Written by Octri · Checked 25 October 2026

What to have to hand

Octri reads Speakeasy’s own files. A .zip of your repository or a single config file, up to 15 MB.

.speakeasy/workflow.yaml
gen.yaml
openapi.yaml

What comes across

Each row is a field the Speakeasy importer reads, and where it lands. Rows marked “By hand” are what is left for you.

What	In Octri	From Speakeasy
Spec	The project’s spec	sources in workflow.yaml
SDK languages	One SDK per target	targets in workflow.yaml
Project name	Project name	sdkClassName in gen.yaml
Package name and author	SDK package settings	The first language block in gen.yaml
Method names	Per-endpoint names	x-speakeasy-name-override
Pagination and streaming	Per-endpoint settings	x-speakeasy-pagination, x-speakeasy-streaming
Overlays and hooks	Applied to the spec, moved as code	By hand

Step by step

Apply your overlays
Overlays change the spec, so apply them first and upload the result. Octri then reads the same document your SDKs are built from today.
Bring it in
Sign up, choose Switch to Octri when onboarding asks for your spec, and drop in a .zip with .speakeasy/workflow.yaml, gen.yaml and the spec. A spec held only in the Speakeasy registry needs your Speakeasy token, so add the file or a public URL to the upload.
Check the preview
Before anything is created, Octri shows what it found: the spec, each target as an SDK language, the SDK and package names, and which endpoints carry x-speakeasy settings. Anything it could not carry is named in a note under the preview, so you know what is left to do by hand.
Set up your SDKs
Check each language in SDK Studio, set the package names past the first one, and move any hook code into the SDK repository, where custom code is kept through each regeneration. Give each language the package name your users already install, and publish the first Octri release as a new major version, since method names and types can change. Every later spec change arrives in the SDK repository as a pull request.
Go live
Publish, then point your custom domain at Octri last, so readers keep the old site until the new one is ready. From then on every merged spec change rebuilds the draft of your docs, SDKs and MCP tools, ready for you to publish.

Questions

The importer skips them and names each one in the preview. Octri generates SDKs in TypeScript, Python, Go, Java, Kotlin, Ruby, PHP, Rust, Swift and Dart.