The Revoye API takes a request from your app and returns the AI response. You send it to https://api.revoye.com/v1/completions with an API key. Revoye queues it, processes it on Revoye Cloud or on your device with Revoye Desk — whichever your plan includes — and returns the response. One OpenAI-compatible API for content, writing, automation and everyday app features.

What you need

A Revoye account with a verified email. Create a Revoye account
A plan that includes Revoye Cloud, Revoye Desk, or both. See plans and pricing
With Revoye Desk only: the app installed on Windows, macOS or Linux, connected to your Revoye account, and signed in to your AI account. Download Revoye Desk
An API key, created in the dashboard. How to create an API key

Revoye Cloud needs nothing installed. If you have all of the above, the quickstart is three commands long.

Start here


Quickstart	Key, status check, first request. About five minutes
How a request is processed	What happens between sending a request and getting the response, on Revoye Cloud and Revoye Desk
Authentication	Creating keys, scopes, rotation, what a key cannot do
Revoye API integration guide	The whole public API in one file, in the dashboard's developer area for plans with developer access. Hand it to your AI coding assistant, or read it yourself

API reference


`POST /v1/completions`	The native endpoint. Everything Revoye can do is expressible here
`POST /v1/chat/completions`	OpenAI-shaped, for existing SDKs
`GET /v1/models`	The models your plan can use, in the shape a model picker expects
`GET /v1/status`	Can a request be processed right now, and how much of your plan is used?
OpenAPI specification	The machine-readable contract

Guides


Retrying safely	Idempotency keys, and why a lost response is not a lost request
Webhooks	Signature verification and at-least-once delivery
Errors	Every code, and whether to retry
Rate limits and capacity	Plan limits, request rate limits, and the difference
Using an OpenAI SDK	One base URL change, and what differs
Best practices	Designing an integration that runs reliably for months

Four things to know before you build

wait: true is a long poll, not a synchronous call. A request is kept safely the moment it is accepted. A dropped connection loses nothing — the response is at GET /v1/completions/{id} or on its way to your webhook.

Most responses arrive in 3 to 90 seconds. Revoye is built for content and automation, not for real-time chat. Design for it: queue work, take webhooks, and do not put Revoye behind a spinner someone is watching.

Your plan sets the limits. Each plan has a monthly token allowance per channel. When it is used up, the API says so with QUOTA_EXCEEDED and tells you when it resets. Rate limits and capacity.

On Revoye Desk, your device has to be online. If Revoye Desk is not running on any of your devices, requests wait in your queue until it is — or move to Revoye Cloud automatically, on a plan that includes both.

Everything is versioned under `/v1`

/v1 is a stable contract. Additive changes — a new optional field, a new error code — may ship at any time. Anything that would break a working integration ships as /v2 alongside it.

Revoye API Documentation