REST API Docs
Fully documented API with OpenAPI specs. Everything you or your coding agent needs to get up and running fast.
Transform video in the cloud with our suite of tools — resize, overlay, AI subtitle, AI voiceover and more
fetch('https://api.bannerbear.com/v5/tools', {
method: 'POST',
headers: { 'Authorization': `Bearer ${API_KEY}` },
body: { subtitles: true }
})Sixteen single-purpose operations for video, images and audio. Each one takes a file URL and returns a file, with no encoding pipeline of your own to run.
Fully documented API with OpenAPI specs. Everything you or your coding agent needs to get up and running fast.
Add template-based image and video generation capability to your Node app. Copy and paste your API key from the dashboard.
Grab an API key from the dashboard and add template-based image and video generation capability to your Python app.
Ruby, PHP, MCP hosted and local. There are many more ways to integrate your tech stack with Bannerbear.
12 month trailing 99.9% uptime. 6 years providing image and video API services. Check our current status.
Trim, crop, resize, join, overlay, filter, soften, subtitle and more, each reachable in a single call.
Transcribe the audio and burn styled captions on, controlling font, colour, outline, shadow and background box.
Break captions where the speaker paused, or into fixed runs of one to ten words each.
Mismatched clips are scaled, padded and re-timed to one resolution and frame rate before being joined.
Clips carrying no audio are given real silence, so joining one never mutes the whole output.
Fonts uploaded to a workspace appear beside the built-in ones wherever text is styled by a tool.
Every job reports progress as it runs, so minutes of video encoding are never an opaque wait.
Convert any rendered video into a looping GIF, capped so the file stays small enough to embed.
Single-purpose media operations you can call directly or chain in a workflow.
Single-purpose media operations that work on files rather than templates — trimming a video, burning subtitles, removing a background, joining clips. Each does one job and returns a file.
Video: trim, crop, resize, join, overlay video, overlay image, apply colour filter, soften, add or replace audio, add cover art, transcribe and burn subtitles, create a GIF preview. Images: remove background, create a video slideshow, create a multi-page PDF. Audio: extract audio, generate a voiceover.
POST to that tool’s endpoint with its inputs. Every tool’s page in the dashboard has an API console showing a working request with the fields filled in.
By URL. Upload the file somewhere reachable, or use a URL produced by an earlier render or another tool.
No. A tool job is queued and reports progress while it runs, because most of them re-encode video. Poll the job or use a webhook rather than waiting on the request.
In credits per job, with the heavier operations costing more — joining videos and building a slideshow cost more than a trim, because they re-encode everything they touch.
The tool transcribes the audio and burns the result onto the video. You control the font, size, colour, outline, shadow, position and background box.
Yes. Leave the language on auto-detect, which works well when the audio is clear, or name the language explicitly if the audio is noisy.
Yes. By default the transcript is broken where the speaker paused. Set words per segment to a number between 1 and 10 to re-cut it into fixed runs instead — useful for the fast, few-words-at-a-time style common on social video.
Set the background to Solid box and choose its colour. Background opacity fades the panel so the video shows through, and outline width controls the padding between the text and the edge of the box.
Yes, the voiceover tool turns text into spoken audio with a natural-sounding voice, which you can then add to a video with the audio tool.
Both cut a subject out. The tool works on a file you give it and returns a file; the Image API version happens inside a render, on a layer of a template.
Yes. The slideshow tool takes images in order, holds each for a set number of seconds, and can cross-fade between them.
Yes. Inputs rarely match, so each one is scaled, padded and re-timed to a common resolution and frame rate before joining. Clips without audio get silence, so the sound of the others is not lost.
1280x720 by default, or whatever width and height you pass. Clips that do not fit are letterboxed rather than stretched.
GIF is an inefficient format and file size grows fast. The caps — 5 seconds, 8fps, 480px wide — keep a preview small enough to load quickly in a page or an email.
Yes, that is what workflows are for. A workflow step can use the output of an earlier step as its input, so trim, subtitle, then add a logo becomes one thing you run.
Yes. Fonts uploaded to your workspace appear in the subtitle tool’s font list alongside the built-in ones.
Start a free trial with 30 API credits, no credit card required
