Make your first BrewReel promo video: run a sample, then use your own product

Understand who does each part

You supply a brief and assets. A language model writes storyboard JSON using the project’s recipes. BrewReel renders the video locally with code. Running an existing sample first helps verify the renderer before introducing model-generated storyboards.

The cards recipe presents features and processes, quiz handles question-and-reveal sequences, and journey follows a character through a scene. Cards and quiz default to 1080×1920; journey defaults to 1080×1350. Check the repository for available recipes and parameters.

Prepare the local environment

The current README lists Node.js 18+ and Python 3.10+, and records testing with Node 22. The DeepSeek Harness plugin has separate runtime requirements. You also need Git, npm and space for dependencies. Follow the project’s x64 requirements on Windows.

1. Install rendering dependencies

git clone https://github.com/Finderchangchang/brewreel.git
cd brewreel/template
npm install
npx remotion browser ensure
cd ..
python -m pip install numpy scipy

If your system only provides python3, use python3 -m pip install numpy scipy for the last line. These commands install rendering components, download a headless browser and install music-generation dependencies. If the browser download fails, inspect network access and the project’s local-browser instructions before changing the storyboard.

2. Validate and render a repository sample

node scripts/validate.mjs examples/ledger.json
node scripts/make.mjs examples/ledger.json --out ../brewreel-out/ledger

Run these from the repository root and keep the output directory outside it. Passing static validation establishes that the storyboard meets validation rules. Complete a render before checking the actual video and delivery files.

The README describes video.mp4, sheet.png, check/, report.txt and manifest.json in the output. Watch the complete video, inspect the report and confirm the manifest corresponds to your storyboard.

3. Replace the sample with your product brief

Specify the product name, audience, core action, features, evidence, assets and acquisition route. Ask the model to follow the project SKILL and the selected industry recipe. Start with the product brief guide. Mark missing facts as unknown instead of letting the model invent performance figures.

Repeat validation, rendering and review. Use validator errors and human observations to revise the storyboard. Model calls and optional voiceover services may incur separate fees.

4. Review before publishing

• Check product names, features, prices and conditions against the brief.
• Check readable subtitles, natural narration and the music-to-voice balance.
• Confirm permissions and labels for real photos, screenshots and illustrations.
• Verify that the closing call to action uses a real acquisition route.
• Apply AI-content labels required by the publishing platform.

Automated checks catch some issues; they do not replace watching the video or reviewing its business claims. A rendered sample is not automatically ready for a customer to publish.

Check both project and dependency licenses

BrewReel uses Apache-2.0. Follow LICENSE and NOTICE requirements when redistributing or modifying the project. Rendered videos do not require project attribution. Remotion has its own license; for-profit organizations with four or more employees need a Company License. See the official terms.

This guide follows the public README. The website update did not independently rerender or benchmark these samples. Check the repository for commands, limitations and version changes.