TwelveLabs

Model

Video understanding models to search inside videos and summarize or query them

Price: 10 free indexing hours, then pay per use
Access: API key (x-api-key header)

About

Video foundation models behind one API. Marengo indexes visuals, speech and on-screen text so text or image queries return clips with timestamps; Pegasus writes summaries, chapters and answers from a video. Pegasus is full in English, partial in 12 more languages. Also on Amazon Bedrock.

What you can do with it

Find the moments in a video library that match a text or image query, with timestamps
Generate summaries, chapters, highlights or answers to questions about a long video
Create video embeddings for your own semantic search or recommendation system

Get started

Create an API key in the TwelveLabs Playground
Create an index with Marengo 3.0, then upload or link your videos
Search the index, or analyze a video with Pegasus

Example

from twelvelabs import TwelveLabs

client = TwelveLabs(api_key="<YOUR_API_KEY>")

results = client.search.query(
    index_id="<YOUR_INDEX_ID>",
    query_text="a player celebrates a goal",
    search_options=["visual", "audio"],
)
for clip in results:
    print(clip.video_id, clip.start, clip.end, clip.rank)

Details

Hosting: Hosted service
Available in: Worldwide
Official SDKs: Python, JavaScript/TypeScript
MCP server: Remote
Works with: AWS