API reference

API reference.

Ompera speaks the OpenAI API. Code that works with OpenAI works with Ompera once you change the base URL, the key and the model.

Base URLhttps://api.ompera.net/v1

AuthorizationBearer YOUR_OMPERA_API_KEY

API keyCreate one in Workspace → API keys.

ContractOpenAPI 3.1 schema for code generators and request validation.

Two kinds of endpoints

ModelsChatTalk to an LLM: yours, one from your pools or the Ompera network. Streaming, tool calls and JSON output. ServicesTranslation, OCR, speech, embeddingsEach has its own endpoint and an engine made for that task. Choose a fixed engine for embeddings.

All endpoints

Models
GET`/v1/models`	The models this key can call right now.
POST`/v1/chat/completions`	Send messages; get an answer or a stream.
Services
GET`/v1/capabilities`	Which services are available right now.
POST`/v1/translations`	Translate text.
POST`/v1/ocr`	Read text from an image.
POST`/v1/audio/transcriptions`	Turn speech into text.
POST`/v1/embeddings`	Turn text into vectors.

There is no /v1/responses or legacy /v1/completions: choose Chat Completions in your client. Errors, retries and limits are the same for every endpoint: see Errors and limits.

Check your key

List the models your key can use. An empty list means no model is online for this key yet.

curl https://api.ompera.net/v1/models \
  -H "Authorization: Bearer $OMPERA_API_KEY"

The examples in this reference read your key from the OMPERA_API_KEY environment variable. Replace MODEL_ADDRESS with an id from this list.

SDKs and frameworks

Any OpenAI-compatible client works: set the base URL and your Ompera key. The official OpenAI SDKs also read OPENAI_BASE_URL and OPENAI_API_KEY from the environment, so existing code can switch without changes.

PythonOpenAI SDK JavaScriptOpenAI SDK FrameworksLangChain, Vercel AI SDK OpenCodeCoding agent config Any appBase URL, key, model OpenAPIGenerate a client

Clients that only speak the Responses API or a provider-specific API need a Chat Completions adapter.