One integration.
Every provider.

infer0 is the OAuth layer for AI inference. Your users bring their own provider keys (OpenAI, Anthropic, Google). You call a single API in whatever SDK format you prefer. We route to their provider.

01 One integration, three SDK formats.
02 Zero key handling. Encrypted at rest.
03 Users pay their own provider bills.
Start building Read the docs

How it works

1

Developer integrates once

Add the infer0 OAuth flow to your app. Use the OpenAI SDK, Anthropic SDK, or Responses API. infer0 translates between formats automatically.

2

User connects their provider

During the OAuth flow, each user picks a provider (OpenAI, Anthropic, Google) and enters their API key. You never see or handle their credentials.

3

App sends requests to infer0

Your app passes the user's access token to /v1/chat/completions, /v1/messages, or /v1/responses. Use the SDK that matches your codebase.

4

infer0 routes to their provider

We look up the user's configured provider and model, forward the request, translate the response, and return it in your requested format. Your app never needs to know which provider is behind the scenes.

For developers

//

SDK of your choice

OpenAI Chat, Anthropic Messages, or Responses API. Use the SDK your codebase already has. No provider-specific routing logic needed.

/**

Zero key management

Provider API keys are encrypted with AES-256-GCM and never exposed to your app. You don't store, handle, or even see them.

*/

Built for streaming

All cross-format streaming is handled server-side. The SDK in your client receives its native stream format regardless of the upstream provider.

For end users

Your account, your keys

Connect your existing OpenAI, Anthropic, or Google account. Your API key stays encrypted and is never shared with the apps you authorize.

Use what you pay for

Already have a subscription or credits? Use them across any app integrated with infer0. No double billing.

Full control

Add, remove, or switch providers anytime. Revoke any app's access with one click from your dashboard.

Early projects using infer0

infer0 is in early development. If you are building something with it, tell us about it and we will feature your project here.

Security & privacy

Prompts stay private

infer0 does not log or store prompt or completion content. Metadata for rate limiting is retained, but message bodies are never written to disk.

Encrypted credentials

Provider API keys use AES-256-GCM before storage. Tokens are hashed. Encryption keys live in Cloudflare's secure infrastructure, separate from the database.

Staff cannot read keys

Encrypted data cannot be read by infer0 staff. The encryption keys are stored in Cloudflare's secure infrastructure, separate from the database.

Revoke anytime

Users can revoke any app's access from their Authorizations page. Revoking immediately invalidates all associated tokens and blocks further requests.

Data retention

Account profiles, encrypted provider configs, and OAuth authorization records are retained. Users can delete providers and revoke authorizations at any time.

If infer0 is unavailable

User API keys with their AI provider remain valid and unaffected. App requests to infer0 will fail until service resumes. Handle this gracefully in your app.