TwelveLabs
TwelveLabs builds video-understanding foundation models (Marengo for search and embeddings, Pegasus for analyzing video and generating text). The TwelveLabs API lets developers upload and index video, run any-to-video semantic search, generate text from video (titles, topics, hashtags, summaries, chapters, highlights, and open-ended analysis), and create multimodal embeddings over a REST interface authenticated with an x-api-key header.
APIs
TwelveLabs Indexes API
Create, list, retrieve, update, and delete video indexes, and manage the indexed videos within them. Each index is configured with the Marengo and/or Pegasus models and the visu...
TwelveLabs Tasks (Upload) API
Upload video by file or public URL and index it into a target index. Create, list, retrieve, and delete asynchronous video indexing tasks and poll their status (validating, pend...
TwelveLabs Search API
Any-to-video semantic search across an index using text and/or media (image) queries, powered by the Marengo model, with visual / audio / transcription search options, clip or v...
TwelveLabs Analyze / Generate API
Generate text from video with the Pegasus model - open-ended analysis via the sync /analyze endpoint (with optional NDJSON streaming), async analysis tasks for longer video, plu...
TwelveLabs Embed API
Create multimodal Marengo embeddings for text, image, audio, and video in one shared vector space - synchronously for short inputs and via asynchronous embedding tasks for longe...
Event Specifications
TwelveLabs Analyze Streaming (HTTP + NDJSON)
AsyncAPI 2.6 description of TwelveLabs' **analyze (text generation) streaming** surface. TwelveLabs does not publish a WebSocket API. The only asynchronous / event-style transpo...
ASYNCAPI