You can make boring history videos for sleep by combining an AI-written script, generated narration, and still images in a video editor. The workflow here uses a short story about Napoleon to show how those pieces fit together.
The sample is smaller than the long videos common in this format, but it demonstrates the full production sequence. It starts with a history topic in Gemini and keeps a consistent narration voice through the finished edit.
Key Takeaways
- Gemini turns the selected history topic into a two- to three-minute sample script, with story sections and visual suggestions for the next production stage.
- ElevenLabs generates the narration one section at a time. The demonstration uses the Roger voice for both audio clips.
- The script’s visual suggestions become prompts for image generation. Those images are downloaded alongside the voiceover files before editing begins.
- CapCut brings the assets together on a timeline. Each image stays on screen for its related narration, and the sequence is previewed before export.
What a Boring History Video Looks Like
Sleepless Historian is the example channel behind this workflow. Its videos present history in a calm, ASMR-style format intended for listening at bedtime. The channel shown publishes more than five times a week.
The production method combines narrated history with supporting images. To demonstrate it, I use a short story about Napoleon and the Bunny Invasion of 1807.
The example channel’s videos are much longer than this sample. The short version shows how to create the individual assets and assemble them before working on a longer timeline.
Choose a History Topic and Write the Script in Gemini
Find an unusual story
I start by searching for funny moments in history. The ideas include the Great Emu War and Napoleon’s encounter with rabbits, and I choose Napoleon and the Bunny Invasion of 1807 for the sample.
That topic becomes the starting point for the script. In Gemini, I enter the selected story and ask for a video script.
Generate the narration and visual suggestions
This stage uses Gemini’s text-generation capabilities to produce the written material. The resulting script runs about two to three minutes and separates the story into sections, including a hook, the setup, and the attack.
The opening describes Napoleon’s military success by 1807. It mentions Austerlitz, his victories over other rulers, and his changes to Europe’s map before introducing the rabbit encounter.
Gemini also supplies visual suggestions alongside the narration. Those suggestions matter later because they provide the starting material for image prompts. The script supplies both the spoken story and ideas for what to show while it plays.
Generate the Voiceover in ElevenLabs
Turn the first script section into audio
Next, I open ElevenLabs and select its text-to-speech feature. I copy the opening section from Gemini into the text area, choose the voice named Roger, and generate the speech.
The first clip reads the passage about Napoleon’s position in Europe in 1807. After listening to it, I download the audio file to my computer.
Then I repeat the process with the next section. That clip begins after the Treaty of Tilsit and introduces Napoleon’s plan to celebrate.
Keep the voice consistent between clips
I use Roger again for the second generation. Switching voices between sections would change the narrator partway through the story, so the demonstration keeps the same voice throughout.
The sequence is straightforward:
- Copy a section of narration from the script into ElevenLabs.
- Generate the speech with the selected voice and listen to the result.
- Download the clip, then repeat with the next section using that same voice.
Each generation uses account credits. The downloaded clips remain separate files until they reach the video editor.
Create Images From the Script’s Visual Suggestions
With the narration downloaded, I return to the visual suggestions in the Gemini script. I copy one of those descriptions and use it as a prompt in the image-generation tool.
After generating the images, I select one and download it. Then I repeat the process with another visual prompt and download a second image.
At this point, the sample has two types of assets ready for editing: the narration clips and the images that will accompany them. The images follow the ideas already attached to the script rather than requiring a separate round of scene planning.
Assemble the Images and Narration in CapCut
Import the files and build the timeline
I use CapCut to put the video together. For a separate workflow using its generation features, our CapCut AI video maker tutorial covers that option. Here, I import the images and voiceover files separately.
After uploading the media, I drag the first image onto the timeline and place the second image after it. Next, I add the first narration clip, followed by the second.
I expand the timeline view so the clips are easier to work with. Then I extend each image’s duration until it covers the corresponding audio.
Match each image’s on-screen duration to its narration clip. The image needs to remain visible until that part of the story finishes.
Preview the sequence before export
Once the images and audio are arranged, I play the timeline from the beginning. The preview starts with Napoleon’s military achievements, introduces the rabbit encounter, and moves into the section about the Treaty of Tilsit.
This playback shows the downloaded assets working together as a video. The narration continues across the image sequence in the order established by the script.
Export the Video After Finishing the Timeline
CapCut includes an export option for the assembled video. Before exporting a full-length history video, the timeline would need more narration, more supporting images, and further editing than this short demonstration contains.
The sample establishes the production sequence: generate the assets, arrange them in order, and adjust the image durations around the audio. A longer video requires carrying that process through the rest of the script.
Frequently Asked Questions
Do I have to use the Roger voice?
Roger is the voice I choose for the demonstration, but ElevenLabs lets you select another voice. The important point in this workflow is to keep your choice consistent across the separate narration clips.
Why do you recommend a paid ElevenLabs plan?
Every speech generation uses credits, which decrease as the clips are created. My recommendation to use a paid plan comes from that repeated voiceover use; the demonstration doesn’t compare plan prices or allowances.
Can I use another editor instead of CapCut?
Clipchamp is another option mentioned for assembling the video. I use CapCut in the demonstration because many viewers are already familiar with it.
Where does Leonardo AI fit into the process?
Leonardo AI is listed among the tools accompanying the tutorial. The image-generation step uses visual descriptions copied from the Gemini script as prompts, then produces images to download for the edit.
Can the finished video be shared directly to TikTok?
CapCut offers direct sharing to TikTok, as shown at the end of the demonstration. That option comes after assembling the media and preparing the video for export.
A Short Sample Shows the Production Sequence
The Napoleon example brings an AI-written history script, a consistent narrator, and generated images into one edited sequence. The key editing task is matching the visuals to the audio.
The demonstration stops at a short sample. A longer sleep-history video follows the same sequence through a longer script and a more complete timeline.
