Chat, images, embeddings and audio on Pendra

Pendra is more than a chat API. The same platform also generates images, creates text embeddings and transcribes audio, all on sovereign UK infrastructure and all behind familiar OpenAI-compatible endpoints. Here is what you can run today.

Chat

Send a prompt, get a response. Pendra serves open chat models through a standard chat completions endpoint, with streaming, tool calling, structured JSON output and vision on the models that support it. If you already use the OpenAI or Anthropic SDKs, you can point them at Pendra and keep your existing code. The chat guide has the details.

Image generation

Generate images from a text prompt through an OpenAI-compatible images endpoint. It suits product mockups, illustrations and anywhere you would otherwise call a hosted image model, with the same promise that your data stays in the UK. See the image generation guide.

Embeddings

Turn text into vectors for search, retrieval and recommendation. Pendra's embeddings endpoint is the building block for semantic search and retrieval-augmented generation, and it runs on the same infrastructure as everything else. See the embeddings guide.

Audio transcription

Upload audio and get back text. Pendra transcribes recordings through an OpenAI-style transcription endpoint, so meeting notes, voice messages and call recordings can be processed without the audio ever leaving the country. See the audio guide.

One API, four capabilities

Everything above sits behind one API key and speaks the same OpenAI-compatible format, so you can mix and match without learning a new interface for each. Browse the documentation to get started, or get in touch if you want to talk through a use case.