PAY FOR GPU MINUTES, NOTHING STORED.
verdantbloom is an OpenAI-compatible API for open-weight AI models, served exactly as their publishers release them: Qwen, Mistral, Google Gemma and others. 25 models today. Open models, one OpenAI-compatible API, billed per GPU minute.
WHAT IT IS
- Publisher releases only. Every model is the publisher's own release (or a standard quantisation of it), under a licence that allows commercial use. The licence is listed next to each model.
- Our own engine. Models run on kinfer, an inference engine we wrote from scratch for NVIDIA GPUs. It serves every text, vision and document model listed on this site.
- Drop-in for OpenAI clients. Point any OpenAI SDK at our base URL. Streaming, images in messages, and a time-to-first-token estimate before you send.
- Nothing you send is stored. We keep meters and money, never the text of requests or replies.
WHO IT IS FOR
Developers and companies building on open models: assistants, document and image understanding, coding tools, data extraction, agents. Not for adult-content services: see the acceptable-use policy.
START IN A MINUTE
1. Create an account with your email (we send a one-time login link). 2. Make a key on your account page. 3. Call the API.
curl https://api.verdantbloom.bar/v1/chat/completions \
-H "Authorization: Bearer $VB_KEY" -H "Content-Type: application/json" \
-d '{"model": "mistral-nemo-12b", "messages": [{"role": "user", "content": "Hello"}]}'