What an AI video analyzer does
An AI video analyzer reads a finished video and returns structured findings instead of a one-paragraph summary. rea-video maps every shot onto the original timeline, labels how the picture was most likely made — live footage, AI generation, 3D, code graphics or screen recordings — and records captions, sound cues and editing patterns as separate layers.
Asking a chatbot vs running a structured analysis
Multimodal assistants such as Gemini can describe a clip you show them, which is a quick way to answer a single question. A dedicated analysis workflow goes further: it processes the full timeline, keeps timestamps for every finding, separates the picture from its caption and audio layers, and produces a report you can re-read, edit and hand to collaborators. For a one-off answer, ask a chatbot; for a production plan you can work from, run the analysis.
What the analysis returns
- A shot table: start time, duration and likely production method for every shot, with representative frames.
- Layer notes: captions and graphics over the picture, voice, music and sound effects, mapped to the moments they occur.
- A production plan: which shots to film, generate or rebuild, which assets you must supply, and per-shot prompts for generation.
Classifications are model-assisted judgments with uncertainty kept visible; they are not proof of the original creator’s tools. Analysis runs locally — this website does not accept uploads.
See what a useful video description includes
Shot 11 of the unboxing case (00:22.900–00:23.567) separates the hand action, fixed camera, text layers and tentative AI classification. The result describes the picture while preserving the reasoning behind it.
How to describe a video with AI · YouTube reference workflow
Is rea-video a free AI video analyzer?
The core is open source: installing and running it locally is free. Your model provider or AI assistant may charge for its own usage.
Can Gemini or ChatGPT analyze a video?
Yes, multimodal chatbots can describe clips and answer questions about them. This workflow organizes that analysis into a timestamped shot table, evidence and a production plan you can revise.
Can I upload my video on this site?
No. reavideo.ai shows examples and guides. The analyzer runs on your machine through Claude Code, Codex or the CLI, and cloud-model data handling depends on your configuration.
What do I need to run it?
Python 3.10+, ffmpeg, and an agent that can run commands. The docs walk through installation step by step.
Which videos can it analyze?
Any local file ffmpeg can read. Use material you have the right to analyze; a breakdown is an analytical aid, not evidence of the original production.
See real analysis examples · Video-to-prompt workflow · Install and analyze your video ↗