Documentation

Documentation

OpenAI- and Anthropic-compatible. Point any SDK at https://api.ultragpt.pro/v1.

API reference OpenAPI
  1. Create an API key and store it as ULTRAGPT_API_KEY.
  2. Add credits to your balance (shared with the UltraGPT app).
  3. Install the SDK (pip install openai / npm i openai) and call a model:
curl https://api.ultragpt.pro/v1/chat/completions \
  -H "Authorization: Bearer $ULTRAGPT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "grok-4.7",
    "messages": [{ "role": "user", "content": "Hello!" }]
  }'

Endpoints

POST/chat/completionsChat completion (OpenAI format), streaming or not.
POST/messagesAnthropic Messages API (Anthropic SDKs, Claude Code).
POST/messages/count_tokensEstimate prompt tokens (Anthropic format).
POST/responsesOpenAI Responses API, with stored conversations.
GET/responses/{id}A stored response (also DELETE, /input_items).
POST/completionsLegacy text completions.
POST/embeddingsText embeddings.
POST/images/generationsGenerate images.
POST/audio/speechText to speech.
POST/audio/transcriptionsSpeech to text (also /audio/translations).
POST/filesUpload a batch input file (also GET, DELETE, /content).
POST/batchesRun up to 5,000 requests asynchronously, outside your rate limits (also GET, /cancel).
POST/rerankRerank documents against a query (Cohere / Voyage / Qwen).
POST/moderationsClassify text and images for harmful content. Free.
POST/videosStart a video generation (async; also GET, GET /content, DELETE).
WS/realtime?model=Realtime speech-to-speech over WebSocket (OpenAI Realtime protocol).
POST/search/{engine}Standalone web search (Tavily, Exa, Perplexity…), per query.
GET/modelsModels your key can use (?type=, ?supported_parameters=, ?input_modalities= filters; also /models/count).
GET/presetsPresets of your workspace (writes need a management key).
GET/keysManagement keys: list, create, update, revoke keys.
GET/activityDaily usage per model and endpoint, last 30 days.
POST/generations/lookupCost and stats of up to 1000 requests at once.
GET/keyThe calling key: limits, spend.
GET/creditsCredit balance and lifetime totals.
GET/generation?id=Tokens, cost and latency of one request.
All docs on one pageNextAuthentication →