- Visible evidence
- At the opening, a woman sits beside a microphone. Around five seconds the camera is closer to her face; the closing shot shows a man seated on a couch.
- Sound / speech
- Not assessed in this annotated layout. Replay the clip to check speech and sound.
- Uncertainty
- The image alone does not establish who is speaking at every moment.
AI Video Analyzer — Understand Your Clip
AI Video Analyzer by Opus 5.5 Video summarizes scenes and speech with timestamps you can replay.
MP4 / MOV / WebM · 1–60 seconds · 12 MiB · up to 1080p Over the limit? Trim or compress first. Rejected files use no credits.
Only Analyze sends your file to the AI provider. Local preview is free.
Conversation: changing speakers · 10s
Report layout example
A conversation shown in successive close and medium shots.
- An illustrative visual note; no model accuracy or transcript claim.
AI Video Analyzer helps editors and creators understand an existing clip through a summary, scene notes, sound observations and approximate timestamps. A report costs 4 credits (about $0.19 at the lowest monthly rate); the 3-credit signup gift does not cover one. Upload an MP4, MOV or WebM lasting 1–60 seconds, up to 1080p and 12 MiB; export watermark-free TXT or JSON notes, with full subtitles and recreation prompts outside this tool’s scope.
How the AI video analyzer works
AI Video Analyzer reviews the selected video and its audio together, then returns a summary and timestamped observations. Choose scene notes, demonstrated steps or your own question; a subtitle file is not required. Replay the cited moments to check what was seen or heard, review uncertain details, and download the report for your editing notes.
Choose your clip
Upload one MP4, MOV or WebM from your device. Trim longer recordings first.
Choose a review focus
Use scene notes, demonstrated steps, or a specific question. The quote appears before analysis.
Review and export
Replay the cited moments, correct uncertain claims, and download the text or JSON report.
Three video analysis report examples
These three annotated layouts show how to separate observations from assumptions: a conversation, a close-up explanation and a workbench demonstration. Each pairs an existing ten-second clip with a short visual note. They illustrate the report format; the sound fields are deliberately unassessed, so visible captions are not presented as verified speech.
Conversation: changing speakers · 10s
At the opening, a woman sits beside a microphone. Around five seconds the camera is closer to her face; the closing shot shows a man seated on a couch.
Close-up: hands and captions · 10s
The man leans forward and moves his hands. A close framing keeps his upper body and face visible. Caption words change across the clip.
Workbench: a visible demonstration · 10s
His hands move across the foreground while the camera remains facing the workbench. Yellow and white captions appear over the lower part of the image.
Does an AI video analyzer read frames or just subtitles?
An AI video analyzer can combine visual details with speech and sound; a transcript summary is limited to words. This tool sends the selected video with its audio for analysis and returns separate visual, audio and uncertainty notes. It does not first extract a subtitle file. Fast actions, small on-screen text and unclear speech can still be missed, so check the cited moments.
| Task | What you get | On this page |
|---|---|---|
| Speech transcription | Spoken words written out in order, sometimes with timed subtitles. | Selected speech notes only; no full transcript or SRT/VTT export. |
| Video content analysis | A summary and observations about scenes, visible actions and sound. | Scene notes, Steps or My question; replayable timestamps and TXT/JSON downloads. |
| Video-to-prompt | A new instruction describing subjects, camera, lighting and motion for a video generator. | No dedicated recreation-prompt output or recovery of the original generation prompt. |
Can an AI video analyzer turn video into a prompt?
Some AI video analyzers turn shot descriptions into prompts for making a similar video. That is a separate output from a content report, and the original generation prompt cannot be recovered reliably from footage alone. On this page, a prompt means the analysis question you provide. Scene notes and Steps supply that question automatically; My question lets you focus the report on one visible action or passage.
Example analysis question — input to the analyzer
Describe the visible steps in order. Give start and end timestamps, separate speech from on-screen text, and mark unclear details instead of guessing.
Is this AI video analyzer free?
AI Video Analyzer reports are paid: a new analysis costs 4 credits, about $0.19 using the lowest monthly credit rate. The 3-credit signup gift does not cover one report. Local file preview and the report examples are free. The smallest credit purchase is $5; downloading a completed report or reopening the same video with the same review settings does not charge again.
Video analysis pricing and limits
AI Video Analyzer charges once for each new video-and-question combination. One report covers a clip up to 60 seconds, 12 MiB and 1080p, including text and JSON downloads. Larger, longer or higher-resolution files are rejected before charging; trim or compress them first. A different question is a new analysis. Technical failures refund the report’s credits; retries require a new confirmation, and completed reports remain free to retrieve.
| Action | Cost | Allowance / output | Watermark |
|---|---|---|---|
| Local preview and examples | $0 | No AI analysis; no account required | None |
| New analysis | 4 credits ≈ $0.19 | One 1–60s file; 12 MiB; up to 1080p | None |
| Saved report and export | $0 | Same video and review settings; text / JSON | None |
4 credits ≈ $0.19 at $0.04714 per credit ($99/month for 2100 credits). Credit purchases start at $5. Signup gift: 3 credits, below one analysis.
See credit plansWhat this video analyzer cannot establish
AI Video Analyzer produces a draft for human review, not a verified transcript or frame-by-frame forensic report. Sampled frames can miss quick actions, and a model can mishear speech or misread small text. The tool cannot establish a person’s identity, intent, health, exact location, copyright permission or audience retention from the video alone.
No platform-link import, full transcript or subtitle export, dedicated recreation prompts, live streams, long recordings, batch review or video editing. The model uses sampled frames, so it cannot check every frame or provide precise sports or music timing. Use a short local copy you are allowed to analyze.
AI video analyzer FAQ
The AI video analyzer works with a file from your device and a signed-in account for paid analysis. These answers cover device use, file handling, specialist feedback and saved results. Choose an excerpt relevant to your question, and retain the original recording so you can verify the report or check a timestamp later.
Can I use this AI video analyzer on my phone?
Yes. Choose a local MP4, MOV or WebM in a modern phone browser; no separate app is required. The file must fit the 60-second, 12 MiB and 1080p limits. Device playback support can vary by codec; MP4 with H.264 video and AAC audio is the practical starting point.
Do I need to sign in, and what happens to my upload?
Sign-in is required for analysis and saved reports. Preview stays in your browser. On Analyze, the file passes through our server to APIMart and the Gemini provider; this feature does not retain a video copy in our storage. The question and report are stored with your account. Provider retention is governed by the provider’s terms; avoid confidential material you cannot share with them.
Can it analyze dance, sports or music videos?
The prompt can ask for visible movements, scene changes or audible events in a short clip. Sampled frames do not support precise timing, joint angles, beat alignment, coaching scores or medical conclusions. The report should be used as viewing notes and checked against the recording.
Can I reopen an analysis after closing the tab?
The latest receipt for your signed-in account is remembered in this browser session. Reopening it retrieves the saved report. Selecting the same file and review settings also recovers an existing result without a new charge. The video itself must be selected again for timestamp playback; download the report to keep your own copy.
Need a short excerpt or timed subtitles? Open the clip editor; caption pricing is separate.
Open clip editor