--- title: "Our first faceless AI video tool: from a written idea to scenes" canonical: "https://blitzreels.com/changelog/creating-without-a-camera" --- # Our first faceless AI video tool: from a written idea to scenes Milestone: 2025-11-25 Article published: 2026-09-26 Status: Early development Our first faceless video tool turned a written idea into a storyboard of 3 to 8 scenes, then generated an image per scene and a narrated voiceover to edit. ![Two AI-generated vertical videos: a pilot in a cockpit and an animated man in a kitchen with captions.](https://blitzreels.com/images/changelog/creating-without-a-camera.jpg) ## Some videos don’t start with footage Until late November 2025, a BlitzReels project usually started with something you had already filmed. That’s fine for a podcast or a vlog. It doesn’t help when the video only exists in your head, like a history explainer or a product walkthrough where nobody wants to be on camera. Faceless channels make these videos every day, with a narrator over images instead of a person talking to the lens. We wanted that kind of video to have a starting point in BlitzReels too, without giving up the edit. ## Write the idea, get a storyboard The first version lived in the media library, behind a Faceless Story button. You described the story in a few lines (between 16 and 1,200 characters) and BlitzReels drafted a storyboard of 3 to 8 scenes, six by default. Each scene came back with a title, a mood, one or two sentences of narration and a visual direction. The storyboard also showed the overall tone and a style guide, so you could read the whole arc before spending anything on images. If your first prompt was thin, Improve prompt rewrote it, and you could accept or reject the new version. And if the arc still didn’t work, you could build the storyboard again before generating a single image. ## From storyboard to scenes and a voice Once the storyboard read well, the rest went like this: 1. Pick the number of scenes and build the storyboard. 2. Read each scene’s narration and visual direction. 3. Choose an image model from OpenAI or Google, and the folder for the results. 4. Generate: BlitzReels creates one image per scene and one voiceover that reads the narration. 5. Open a project and place the images and the voiceover on the timeline like any other media. The scene prompts deliberately asked for no faces and no readable text. Images were saved as reusable media, so a shot made for one story could come back as B-roll in another edit. From there, the [video editor timeline](https://blitzreels.com/features/pro-timeline) is where you trimmed, reordered and timed each image to the voice. The editor’s zoom and mask effects got an update the same week, handy for giving a still image some movement. ## What you could make with it Take a one-minute explainer on why the Titanic sank. You write the idea, get six scenes (the launch, the ignored warnings, the iceberg, the lifeboats…), read the narration, then generate. Six images and one narration track land in your media library, ready to arrange. The same flow suited a product demo that never shows a face (the example in the prompt box) or a quick list video for a faceless channel. The pacing and the final order were still your call, in the editor. One tip: read the narration out loud before you generate. Each scene only had a sentence or two, so a weak line stood out right away, and rebuilding the storyboard was quicker than reworking the edit afterwards. ## An early version, and what came after This was early development, and it showed. Scenes were still images, there was a single default narrator voice, no music or sound effects, and nothing to keep a recurring character looking the same from one scene to the next. Most of that changed over the following months. [Background music and sound effects](https://blitzreels.com/changelog/give-your-story-a-soundtrack) joined in January 2026, and [reusable characters](https://blitzreels.com/changelog/characters-that-belong-to-the-same-story) arrived in August. Today the [faceless video generator](https://blitzreels.com/faceless-video-generation) sits with [every other AI tool under Generate](https://blitzreels.com/changelog/every-ai-tool-in-one-place), on a paid plan or the no-card trial, and each generation uses credits. It replaced the Faceless Story button from this first version. ## Questions ### What is a faceless AI video? A video where nobody appears on camera: a narrator speaks over generated visuals. In BlitzReels it starts from a written idea or script instead of footage. ### Did the first version produce a finished video? No. It generated one image per scene and a voiceover of the narration, saved to your media library. You assembled and timed the video yourself in the editor. ### How many scenes could a storyboard have? Between three and eight, with six by default. Each scene had a title, a mood, a line or two of narration and a visual direction. ### How do I make a faceless video in BlitzReels today? Open Faceless video under Generate in the dashboard, describe your topic and start the generation. It needs a paid plan or the no-card trial, and each generation uses credits. ## Start a video without filming it See how faceless video generation works today, from a topic to a narrated video you can keep editing. [Explore faceless video](https://blitzreels.com/faceless-video-generation) We wrote this story later, looking back at our product history. The date at the top marks the work described here; features and access may have changed since then.