The request names the Cloud Video Intelligence API, which is Google Cloud's product; its page 1 lists it as an official API, and its action record 5 shows asynchronous video annotation. Alternatives that analyze video content are FrameFetch 3 and Veedcrawl 4, which focus on social-video URLs, and Replicate's CogVLM2 video model 2 for visual understanding.
- 1
Google Cloud Video Intelligence APIgoogle.comverified and signed
Runs asynchronous video annotation through the videointelligence.videos.annotate action, with results retrieved via the long-running operations interface 5.
official:API · Cloud Video Intelligence API
for agents: videointelligence.videos.annotate · videointelligence.projects.locations.operations.list · videointelligence.operations.projects.locations.operations.cancel
- 2
FrameFetchframefetch.netverified and signed
Extracts metadata, transcript, sampled frames and on-screen text from one public social-video URL via the framefetch_extract action 7.
+1 community tools built for it
for agents: framefetch_extract · framefetch_platform_capabilities
- 3
Replicate CogVLM2replicate.comverified and signed
Runs the CogVLM2 visual language model for image and video understanding through the chenxwh/cogvlm2-video action 6.
for agents: chenxwh/cogvlm2-video · chenxwh/cogvlm2
- 4
Veedcrawlveedcrawl.comwebsite read, tools not yet checked
Provides transcripts, metadata and structured extraction for public videos from YouTube, TikTok, Instagram, X and Facebook through an MCP server and API 4.
Sources
- 1AI and Cloud Computing Services | Google Cloudcloud.google.com · read 2026-10-29
- 2Replicate - Run AI with an APIreplicate.com · read 2026-10-29
- 3FrameFetch — Video Data API for AI Agents | Transcript, Frames, On-Screen Textframefetch.net · read 2026-10-29
- 4Veedcrawl - Video intelligence API for AI agentsveedcrawl.com · read 2026-10-29
- 5Cloud Video Intelligence API: declares "videointelligence.videos.annotate" (the record we read and signed)unfragile.ai · read 2026-10-24
- 6chenxwh/cogvlm2-video: declares "chenxwh/cogvlm2-video" (the record we read and signed)unfragile.ai · read 2026-10-24
- 7framefetch: declares "framefetch_extract" (the record we read and signed)unfragile.ai · read 2026-10-24
For agents: the 10 actions behind this answer
videointelligence.videos.annotate · Cloud Video Intelligence API verified and signed
Performs asynchronous video annotation. Progress and results can be retrieved through the `google.longrunning.Operations` interface. `Operation.metadata` contains `AnnotateVideoProgress` (progress). `
chenxwh/cogvlm2-video · chenxwh/cogvlm2-video verified and signed
CogVLM2: Visual Language Models for Image and Video Understanding
xai/grok-imagine-video-extension · xai/grok-imagine-video-extension verified and signed
Extend videos with xAI's Grok Imagine Video model. Provide a source video and describe what happens next.
nicolascoutureau/video-utils · nicolascoutureau/video-utils verified and signed
google/veo-3.1 · google/veo-3.1 verified and signed
New and improved version of Veo 3, with higher-fidelity video, context-aware audio, reference image and last frame support
lucataco/videollama3-7b · lucataco/videollama3-7b verified and signed
VideoLLaMA 3: Frontier Multimodal Foundation Models for Video Understanding
kwaivgi/kling-v3-video · kwaivgi/kling-v3-video verified and signed
Kling Video 3.0: Generate cinematic videos up to 15 seconds with multi-shot control, native audio, and improved consistency
turian/insanely-fast-whisper-with-video · turian/insanely-fast-whisper-with-video verified and signed
whisper-large-v3, incredibly fast, with video transcription
jd7h/propainter · jd7h/propainter verified and signed
Object removal, video completion and video outpainting
kwaivgi/kling-v3-omni-video · kwaivgi/kling-v3-omni-video verified and signed
Kling Video 3.0 Omni: Unified multimodal video generation with reference images, video editing, native audio, and multi-shot control
Labels are the instruments' verdict and are never changed by ranking. Agents and robots get this same answer as JSON at /api/v1/find?q=…&answer=1 and through the MCP server.