Inference API
Open-model inference on our own infrastructure in Spain
Serenity Edge is an OpenAI-compatible API that serves open models and the Orion family from Substrate AI’s GPUs in Valencia. Zero data retention. Data residency in the European Union.
Change the base URL. The rest of your code stays the same.
curl https://api.serenityedge.ai/v1/chat/completions \
-H "Authorization: Bearer $SERENITY_EDGE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "qwen3.6-35b-a3b",
"messages": [{"role": "user", "content": "Hello"}],
"stream": true
}'Any OpenAI-compatible SDK or tool works with Serenity Edge.
- 131,072-token context window
- Up to 32,768 output tokens
- Models served in NVFP4 precision
- Priced per million tokens, in USD
The essentials
OpenAI compatible
Chat completions with streaming, tool calling, structured outputs and prefix caching. Works with the SDKs and tools you already use.
Your data stays in Spain
Requests are processed on Substrate AI’s own infrastructure in Valencia. Prompts and completions are never written to disk or logs; prompt prefixes may stay in a memory cache for less than 24 hours and are never used for training.
Two families, one deployment
Open models such as Qwen, and Orion, Serenity’s own model family, served on the same infrastructure through the same API.
Three ways in
Through OpenRouter, by selecting the Serenity Edge provider.
With a Serenity Star account: API keys and credits.
Under contract, for organisations that need specific terms and an SLA.