Unified AI API for LLMs, Images, Video & Music | OmniaKey

# OmniaKey > Unified AI API for LLM, image, video and music models. One prepaid key connects agents and apps to multiple providers with usage-based billing. Requests use the model ID you select, without silent substitution. OmniaKey is a unified AI API platform. Text clients choose the documented OpenAI-compatible, Anthropic-native or Gemini-native API; image, video and music use asynchronous media tasks. Model availability, supported protocols, billing units and current prices are maintained in the catalog and docs. ## Endpoints - OpenAI-compatible: `https://api.omniakey.com/v1` - Anthropic-compatible: `https://api.omniakey.com` (SDK appends `/v1/messages`) - Gemini-compatible: `https://api.omniakey.com/v1beta` ## Public API scope - OpenAI-compatible Chat Completions: `POST /v1/chat/completions` - OpenAI-compatible Responses: `POST /v1/responses` - OpenAI-compatible Responses compaction: `POST /v1/responses/compact` - OpenAI-compatible model discovery: `GET /v1/models` - Anthropic Messages: `POST /v1/messages` - Gemini Generate Content: `POST /v1beta/models/{model}:generateContent` - Gemini model discovery: `GET /v1beta/models` - Media tasks (image, video, music): `POST /v1/media/tasks` (requires an `Idempotency-Key` header), poll `GET /v1/media/tasks/{id}`, cancel `POST /v1/media/tasks/{id}/cancel`; finished assets are returned as signed URLs Do not state that OmniaKey publicly supports embeddings, rerank, files, fine-tuning, or batches unless those capabilities appear in the docs and dashboard catalog. ## Canonical positioning - Category: unified AI API platform — one prepaid key for LLM, image, video and music models, with text protocols and asynchronous media tasks - Current public catalog: curated Claude, GPT, Gemini, Grok, Qwen, DeepSeek, GLM, Kimi, MiniMax, MiMo, Hunyuan, and Jev models, plus image, video, and music models served as asynchronous media tasks - Primary users: developers using Claude Code, Cursor, Codex, Cline, aider, OpenAI-compatib