AI B-Roll Editor
The AI B-Roll Editor is an AgentReel tool that cuts B-roll to a voiceover using only your own footage. You upload your clips and either a script or your own recorded voiceover. AI reads every clip, picks a shot for each sentence, times it to the voice, and renders one MP4 with burned-in captions and music. It works in English and Chinese. It is built for YouTubers, faceless channel creators, marketers and anyone who has more footage than time to edit it.
Is it free? A new account gets one free B-roll video: up to 60 seconds of script, up to 50 clips, one AI voice generation or your own recording, 0 credits. After that the editor charges step by step and shows the price before each step.
Quick answers
Does it use only my own footage?
Yes. The AI B-Roll Editor picks every shot from the clips and photos you upload. It has no stock footage library and generates no AI shots, so nothing in the video comes from outside your project. When no clip fits a sentence, the editor uses the closest shot from your own footage and lists that sentence under "substitute footage" before you render, so you can add a better clip or a director note and match again.
Can I use my own recorded voiceover?
Yes. Upload your recording as MP3, M4A or WAV, up to 5 GB and 8 minutes, and the editor keeps your voice exactly as recorded. You do not need a script: the editor transcribes the recording, and the transcript becomes the script that the footage is matched to. You can fix any word in the transcript afterwards; that only changes the captions. The transcription costs 1 to 3 credits by the length of the recording, and it is free in the free trial.
Does it pick a shot for every sentence automatically?
Yes. The editor splits your script or transcript into phrases, reads every clip, and picks the footage that matches what each phrase says. It also searches your clips for the people, places and products that the script names. Each shot is timed to the voice. You check the result phrase by phrase on the timeline before you render, and you can steer the choice with director notes such as "use the drone shots for the opening".
Does it support Chinese?
Yes. The script can be in Chinese or English, a Chinese recording is transcribed and matched like an English one, and the captions are burned in in Chinese. The 6 preset AI voices are English, so for Chinese narration upload your own recording or use a cloned voice.
Does it add subtitles?
Yes. Every video gets burned-in captions synced to the voice, in one of 7 styles. With your own recording, the captions follow the words you actually say. You can also export the timeline for Premiere Pro, DaVinci Resolve or Final Cut Pro, with the captions as an SRT file.
What you can make
- A YouTube video with voiceover from your travel, vlog or event footage.
- A vertical Shorts, Reels or TikTok video from the same clips, in 9:16.
- A faceless explainer: your script, read by an AI voice, over your own footage.
- A product or tutorial video that mixes footage with screenshots of your app or website.
- A video in your own voice, from a cloned voice or a recording you upload.
- A photo story: photos become 4 second slow push-in shots between your clips.
How it works
- Open the editor. Go to AI B-Roll Editor and start a new project.
- Add your footage. Drop in videos and photos, in any order. You do not need to sort or rename them. Uploads keep running while you browse other pages, resume after a reload, and skip files that already finished.
- Add your narration. Paste a script in English or Chinese, up to 8000 characters (about 8 minutes of speech). Or upload your own recording instead: the editor transcribes it, and you do not need a script.
- Add director notes (optional). Tell the editor what you want, in plain words: which shots to favor, which to avoid, the mood. Up to 8000 characters.
- Pick a voice. With a script, choose one of 6 preset English AI voices or a voice you cloned. With your own recording, your voice is kept as recorded and the captions follow your words.
- Pick music, captions and pace. Choose a music preset or upload your own track. Choose a caption style and a pace (fast, medium or slow). Choose 16:9 or 9:16.
- Let AI pick the shots. AI reads every clip and picks footage from your clips for each phrase of the script.
- Check the timeline. Go through the timeline phrase by phrase before you render. Add a screenshot overlay where a phrase needs one.
- Render. The editor renders one MP4 with voiceover and burned-in captions. You can close the tab: the cut keeps running and we email you when it is ready. If it stops, the email says why, and the credits for the failed step come back.
- Download or share. The finished video is in My Works.
Options and limits
| Item | Limit / options |
|---|---|
| Items per project | No limit |
| Video files | .mp4, .mov, .m4v, no size limit |
| Photos | .jpg, .png, .webp, .heic; each becomes a 4 second slow push-in shot |
| Upload size | No limit per project or per day |
| Script | Up to 8000 characters (about 8 minutes of speech), English or Chinese |
| Director notes | Up to 8000 characters |
| Output | MP4, 16:9 (1920x1080) or 9:16 (1080x1920), with voiceover and burned-in captions |
| Caption styles | Box, outline, thick outline, shadow, yellow, label, rounded |
| Pace | Fast, medium, slow |
| Preset voices | 6 English AI voices |
| Voice cloning | Paid accounts, up to 3 voices. Sample of 30 seconds to 3 minutes, MP3, M4A, WAV or WebM, up to 40 MB. Consent of the speaker required |
| Your own recording | MP3, M4A or WAV, up to 5 GB and 8 minutes; transcribed, no script needed |
| Footage source | Only your own uploaded clips and photos; no stock library, no AI-generated shots |
| Captions | Burned in, synced to the voice; SRT file in the timeline export |
| Timeline export | Premiere Pro, DaVinci Resolve, Final Cut Pro |
| Music | Off, Upbeat, Warm, Documentary, Tech (free presets), or upload MP3, M4A or WAV up to 5 GB |
| Screenshot overlays | PNG or JPEG up to 10 MB, up to 8 per phrase, 40 per project |
| New projects | Up to 20 a day |
| Storage | Footage and results stay until you delete the project |
Every new voice you generate saves a version. Switching back to an earlier version is free.
There is no limit on the number or size of your clips. The web editor uploads every file first, so hundreds of files or big 4K footage take a while in the browser. For large libraries we recommend the free desktop app for Mac and Windows: it works with the footage on your computer without uploading it, and renders at no render cost.
Free trial
A registered account that has never purchased gets one free B-roll video. The trial costs 0 credits and has these bounds:
| Trial bound | Limit |
|---|---|
| Number of trial videos | One per person |
| Script length | Up to 60 seconds |
| Clips | Up to 50 |
| Voice | One AI voice generation, or your own uploaded recording |
The same script with the same voice again stays free. A new AI voice after the first one needs credits, unless you upload your own recording. If your project goes past the bounds, the editor shows you how to continue with credits.
Pricing
After the trial, the editor charges step by step. Before each step it shows the price and asks you to confirm. A step that fails is refunded automatically.
| Step | Typical cost |
|---|---|
| Reading your clips | About 1 credit per 8 clips |
| Picking shots for your script | About 7 to 34 credits per 30 seconds of script, depending on how many clips you have |
| Render | About 2 credits per minute, plus 1 |
| AI voice | Same as narration: 22 credits per 1000 characters |
| Transcribing your own recording | 1 to 3 credits by the length of the recording (8 minutes at most) |
| Music presets | Free |
These are typical numbers. The editor shows the exact price before each step.
Invite credits can pay every B-roll step except the AI voice. Voice cloning needs a paid account. See Credits, plans and free usage for plans and packs.
Tips for better results
- Write for the ear. Short sentences give the editor clear phrases to match shots to.
- Add more footage than you need. More clips give AI more choice for each phrase. Matching costs a little more with more clips, so drop clips that are clearly unusable.
- Use director notes. A line like "use the drone shots for the opening, avoid shaky clips" helps the shot choice.
- Check the timeline before you render. Going phrase by phrase is the cheapest place to fix a shot.
- Use screenshots for anything on a screen. For an app or website, a screenshot overlay is sharper than filmed footage of a screen.
- Record your own voice for a personal channel. Upload your recording and the captions follow your words.
FAQ
Do I need to sort or trim my clips first?
No. Upload them in any order. AI reads every clip and picks the part that fits each phrase.
Can I make a Chinese video?
Yes. Your script can be English or Chinese, and a Chinese recording is transcribed and matched like an English one. The 6 preset voices are English, so for a Chinese voice use a cloned voice or upload your own recording.
Do I need a script if I upload my own recording?
No. The editor transcribes the recording and matches the footage to what you say. You can correct the transcript; that only changes the captions.
What if no clip fits a sentence?
The editor uses the closest shot from your own footage and lists the sentence under "substitute footage". Add a clip that fits, or a director note, and match again before you render.
What if I do not like a shot?
Check the timeline phrase by phrase before you render. That is the place to catch a shot you do not like, before you pay for the render.
What happens if the cut fails?
We email you why it stopped, and the credits for the failed step come back automatically.
Can I close the browser while it renders?
Yes. The cut keeps running and we email you when it is ready.
Who owns the video?
You keep all rights to the footage you upload. You can use the videos you create, subject to the terms of service. See Privacy, rights and safe use.
How long are my files kept?
Footage and results stay until you delete the project.