Embedding Inference API

Turn images into embeddings.

An OpenAI-compatible embedding service for Precision AI models. Create an API key, explore embeddings right in your browser, or call the API directly from your own code.

Quickstart

Copy-paste curl + Python recipes for embedding text, images, dense patch tokens, and similarity - each with one-click Open in Explorer.

Read the Quickstart

Portal

Create and manage your sk-pai- API keys, track your usage and quota, and (for admins) manage users and tiers.

Open the Portal

Explorer

Upload one image or a batch, generate embedding vectors, and inspect them — dimensions, L2 norm, a value heatmap, and pairwise similarity — all in your browser.

Open the Explorer

AgriViz

Explore agricultural image datasets visually — interactive 3D embedding maps, parallel coordinates, charts, and per-image previews, all in your browser.

Open AgriViz

API docs

The interactive OpenAPI reference at /docs. Point the OpenAI SDK at this server's /v1 base URL with your key and call embeddings.create().

Open the API docs

Get started

1

Create a key

Sign in to the Portal with your Precision AI account (email & password), then create your own sk-pai- API key.

2

Explore in your browser

Open the Explorer, sign in with your Precision AI account, and drop in images to generate embeddings interactively.

3

Or call the API

Send the key as a bearer token to /v1/embeddings. Start with the Quickstart recipes, or see the API docs for the full reference.