Introduction

The Tengr.ai Public API gives you the Tengr.ai platform's image and language stack as a set of REST endpoints. The same models that drive Tengr.ai's apps are available here without a UI in the way.

What you can do

Generate images from a prompt. Pick one of our trained model families. Selena is the workhorse for high-fidelity creative work. Quantum and Hamilton trade off speed and detail differently. Each family ships with its own catalogue of styles and aspect ratios.

Vary or edit images you already have. Drop in a reference image and ask for a variation through Selena Variete, or edit one directly with Photo Pilot. Face refiner cleans up portrait detail, Face swap replaces a face from a source image.

Run utilities on images. Upscale a small image, push fidelity further with Detailer Upscaler, remove the background, generate a segmentation mask, or pull a depth map out of a single picture.

Keep a consistent visual identity. Style-DNA lets you encode a style once and reuse it across generations, so a series of outputs feels like it came from the same place.

Use the Textcraft LLM. For text-only work, call into our LLM through a single endpoint. Useful for prompt rewriting, copy generation, or anything you'd reach for a chat model to do.

Authentication

Every endpoint uses API-key auth via the standard Authorization: Bearer ... header. Your key carries two access dimensions, both enforced on every request.

Scope sets your tier of capability. Three levels in strict hierarchy: admin > manager > generator. Each endpoint declares the minimum scope it needs.

Domains set which feature areas the key can reach: artworks, style-dna, textcraft, or * for all of them.

A manager-scope key with domains: ["artworks"] can still only hit /v2/artworks/*, even though its scope is high enough for other domains.

Ask your organization's admin for a key, or have them provision one through POST /v2/organizations/:orgId/api-keys.

Webhooks

Artwork creation in V2 is asynchronous. When you POST to a generation endpoint you get a 202 back immediately with an artwork ID, and the result lands at the webhook URL you supplied with the request once generation finishes.

Outbound webhooks are signed with HMAC-SHA256 using your organization's webhook secret. Each delivery carries a Tengrai-Timestamp and Tengrai-Signature header. Verify against ${timestamp}.${rawBody} before trusting the payload. The secret is rotated through POST /v2/organizations/:orgId/webhook-secret.

Versioning

V2 is the current surface. This reference describes V2 only.

V1 is frozen at v1.4.0. It stays available during the transition window but is not actively maintained. Its reference lives under V1 (deprecated) in the version dropdown at the top of this page. Use V2 for new work.

Contact

Tengr.ai Support Team: support@tengrai.com