Skip to content

Video Background Remover API

Use your credits to remove video backgrounds in AI assistants and scripts.

Remove video backgrounds with an API

Send a public HTTPS video URL or upload a file up to 500 MB and 300 seconds long. This video background removal API runs asynchronously: estimate the credits, create a job, then poll its status for download links. Results stay available for 7 days.

Choose transparent WebM with audio (webm), HEVC MOV (hevc_mov), ProRes 4444 (prores), green screen MP4 (green_mp4), replacement-background MP4 (bg_mp4), or a PNG sequence ZIP (zip).

Connect your AI assistant (MCP)

Use this MCP server address with OAuth. Sign in and allow your assistant to use your credits.

Install in Claude Code

  1. Install the server

    Run this in your terminal.

    claude mcp add --transport http vbr https://videobackgroundremover.ai/mcp

  2. Sign in when your assistant opens the browser

    Run /mcp in Claude Code to sign in.

  3. Try it

    Remove the background from my video at <your video URL> and return a transparent WebM.

Or let your assistant install it

Copy this prompt and paste it into your assistant for guided setup.

Show prompt

Install the Video Background Remover MCP server for me: run `claude mcp add --transport http vbr https://videobackgroundremover.ai/mcp`, then tell me to run /mcp to sign in. After I sign in, call get_account to confirm it works and show my credit balance.

estimate_cost → remove_background → get_job

Estimate first, start the job, then poll every 10–20 seconds for download links.

API keys

Checking your account…

Connected apps

Checking your account…

Quickstart

Read the OpenAPI 3.1 specification for request and response schemas.

Send your API key as a Bearer token. Replace the video URL with a public HTTPS video and use the returned job ID to poll.

export VBR_API_KEY='YOUR_API_KEY'

# Estimate credits (does not start a job)
curl https://videobackgroundremover.ai/api/jobs/estimate \
  -H "Authorization: Bearer $VBR_API_KEY" \
  -H 'Content-Type: application/json' \
  -d '{"seconds":10}'

# Start background removal
curl https://videobackgroundremover.ai/api/jobs \
  -H "Authorization: Bearer $VBR_API_KEY" \
  -H 'Content-Type: application/json' \
  -d '{"video_url":"https://your-domain.com/video.mp4","seconds":10,"format":"webm"}'

# Poll with the returned job ID every 10–20 seconds
curl https://videobackgroundremover.ai/api/jobs/JOB_ID \
  -H "Authorization: Bearer $VBR_API_KEY"

Green screen and transparent video output

For a green screen API workflow, choose green_mp4: it places the subject on solid green for chroma keying in editing software. Choose webm or prores for an alpha channel instead, so you can composite the subject over another video without keying out a color. WebM works for web playback; ProRes 4444 suits editing workflows.

Pricing

HD processing costs 2 credits per second up to 30 fps, with a 5-second minimum. Higher frame rates use more credits. The initial reservation is settled against the actual duration and frame count: excess credits are refunded, and any shortfall is charged. API and MCP use paid credits only.

See credit packs and pricing →

FAQ

Is there a free API tier?

No. API and MCP requests use paid credits only. You can try a free preview on the website before buying credits.

What video length and size are supported?

Each input video can be up to 300 seconds (5 minutes) and 500 MB. Import a public HTTPS video URL or upload a file. URL imports must provide Content-Length; URLs on private networks are not supported.

Does the output keep audio?

Video exports keep the original audio when the source has an audio track, including transparent WebM and MOV files and MP4 exports. The PNG sequence ZIP contains image frames, so it has no audio track.

How long are results kept?

Results are kept for 7 days after processing. Download URLs expire after 24 hours; request fresh links with get_job or GET /api/jobs/JOB_ID while the result is still available.

API key or OAuth — which should I use?

Use an API key for scripts and backend integrations, and keep it out of browser code and public repositories. Use OAuth when connecting an MCP assistant: it signs in to your account without receiving your API key. You can revoke keys and disconnect apps on this page.

Can I use it from Python?

Yes. Use Python requests with an Authorization: Bearer YOUR_API_KEY header. Call requests.post() with json={"seconds": 10} on /api/jobs/estimate, then POST a video_url, seconds and format to /api/jobs. Read the returned job.id and call requests.get() on /api/jobs/JOB_ID every 10–20 seconds until it is ready, then download the result.