Creating Manual Projects (AI Video Creator)

While the Autopilot feature is fantastic for background growth, there will be times when you want to take granular control over your content. Perhaps breaking news just dropped, or you have a highly specific idea that you want to manually oversee.

The Manual Projects tab is your personal AI Video Creator studio, offering powerful Text-to-Video generation capabilities with extreme customization.

Step 1: Prompting the AI

Navigate to the New Project tab. Here, you start with the foundation of your video: the Idea.

  1. The Topic/Prompt Box: Be as specific as possible. Instead of typing "Space facts," type "A terrifying fact about black holes and spaghettification, ending with a call to action to subscribe." Our specialized prompt-engineering backend will translate this into a high-retention script.
  2. Video Format: Choose between Shorts (9:16) for YouTube Shorts, TikTok, and Instagram Reels, or Long Form (16:9) for traditional YouTube videos.
  3. Target Audience/Tone: You can specify if the tone should be educational, scary, funny, or motivational. This data is fed directly into the script generation phase.

Step 2: The Script & Voice Generation

Once you hit "Generate Script," CinematicAI connects to OpenAI to write a script optimized for viewer retention (utilizing hooks, body content, and strong conclusions).

  • Script Editor: You have full access to edit the generated script. Pro Tip: Keep your sentences relatively short. This makes the AI-generated subtitles punchier and easier to read on mobile devices.
  • Voice Selection (ElevenLabs): Choose your narrator. The voice you choose drastically impacts the mood of the video. For example, use a deep, raspy voice for true crime, or an energetic, bright voice for top 10 lists.

[!WARNING] If you make manual edits to the script after generating the audio, you must regenerate the audio track to reflect your new text!

Step 3: Visuals & B-Roll (Text-to-Video Generation)

A great Faceless YouTube Channel relies heavily on visual storytelling. CinematicAI handles visual generation in two ways depending on your settings:

  1. Stock Footage Matching: The AI analyzes your script sentence-by-sentence and queries premium stock libraries (like Pexels or Storyblocks) to find highly relevant B-roll clips.
  2. AI Image Generation: For highly abstract or specific topics (like "Cyberpunk cityscapes"), CinematicAI can use Stable Diffusion or Midjourney APIs to generate custom, unique images that are slowly panned (Ken Burns effect) to create motion.

Step 4: Rendering and Dynamic Subtitles

The final step of manual video editing with AI is the render. CinematicAI stitches together the audio, the visuals, and automatically applies dynamic subtitles.

  • Dynamic Subtitles: In the era of TikTok and Shorts, hardcoded subtitles are mandatory. CinematicAI automatically highlights the active word being spoken, adding pop-in animations and emojis to retain maximum viewer attention.

Once the rendering is 100% complete, you can download the .mp4 file directly or use the one-click "Publish to YouTube" button.


Frequently Asked Questions (FAQ)

Q: Can I upload my own script instead of using the AI generator? A: Yes! In the Manual Projects tab, simply bypass the prompt generation step and paste your pre-written script directly into the Script Editor. The system will take it from there.

Q: How long does it take to render a video? A: A typical 60-second YouTube Short takes between 2 to 4 minutes to fully generate the script, audio, source the visuals, and render the final HD .mp4 file. Long-form videos (5-10 minutes) may take 15-20 minutes.

Q: Can I change the style or color of the automatic subtitles? A: Yes. Before rendering, click on "Subtitle Settings." You can choose from various popular styles (e.g., the "Hormozi style", bold yellow text, typewriter effect) and adjust the font size to your liking.

Q: What happens if the AI picks a weird stock video for a scene? A: CinematicAI includes a "Scene Editor." Before you hit the final render button, you can review the timeline. If a specific clip doesn't fit the context, you can click on it and ask the AI to "Replace Clip" or manually search for a better keyword.