Free AI Workflow: Script to Viral YouTube Video in Hours

Proof of Concept: Stick Figures Hit 7.5M Views

A YouTube channel launched 2 months ago with only 14 videos of hand-drawn stick figures (no animation, no video clips, just still images) accumulated 7.5 million total views and 137,000 subscribers. This demonstrates that viral success doesn't require expensive tools or design skills—only the right workflow and rhythm.

Stick-figure channel metrics in first 2 months

Why Most Creators Fail: The Wrong Workflow Order

The standard workflow (script → scenes → voiceover → sync) creates a collision between unrelated elements. When scenes are generated first and voiceover second, they lack natural rhythm, causing viewers to click away even if they can't articulate why. The correct order is voiceover first, then scenes—always.

Workflow sequence determines video rhythm and viewer engagement

One Prompt Generates Five Viral Topic Ideas

Paste a single engineered prompt into Claude AI's free plan, and it returns five pre-optimized viral topic ideas. Each idea is designed to perform well if the full workflow is followed. Select one, and Claude generates a complete long-form video script (5–15 minutes) ready to use, plus a downloadable text file.

5

Viral topic ideas generated per prompt

Claude AI free plan output from single engineered prompt

Generate Natural-Sounding Voiceover on 11Labs Free Plan

Copy your script, visit 11Labs free plan (10,000 credits monthly), select the 'Raunak' voice (described as viral, relatable, and real), paste the script into text-to-speech, and download the audio file. This voiceover becomes the foundation for all scene timing.

Transcribe Audio with Exact Pause Timestamps Using Faziscribe

Upload the voiceover to Faziscribe.ai (free account: 10 minutes of AI credit; create new free accounts to continue). Select 'Accuracy' mode, transcribe, and the tool identifies every pause with exact frame-level timestamps (e.g., pause at 1 second 12 frames, 3 seconds 18 frames). This eliminates guessing where scenes begin and end. Works in any language.

Faziscribe workflow: from audio to precise scene timing

Claude Generates Text-to-Image Prompts for Every Timestamped Line

Paste the full timestamp script from Faziscribe back into Claude AI. Claude generates a detailed text-to-image prompt for every single timestamped line. It works in batches; type 'next' to continue until all lines have prompts. Then type 'yes' to download a complete prompt file with every scene, timestamp, and prompt in one text file.

Manual Image Generation vs. Bulk Automation

Generating 100+ images manually in Google Imagen would take hours. Instead, use the free Zappy Flow Chrome extension (built by the creator, developer name: zappywala.ai) to automate bulk generation. Set up Google Imagen with Nano Banana 2 model (best results without hitting limits), paste all prompts with line breaks, and the extension generates and downloads all images automatically in the background.

Time savings from automation: manual vs. bulk generation

Zappy Extension Setup and Execution

Install Zappy Flow extension (verify developer: zappywala.ai), open Google Imagen, turn off Agent Mode, set image mode to 16:9 aspect ratio with output of 1 per prompt and Nano Banana 2 model. Paste all prompts (one line break between each) or upload the prompt text file. Toggle off 'include serial number in filename,' set download folder, and hit generate. The extension runs in the background—you can switch tabs or minimize the window. Do not close the tab, browser, or extension until complete.

103

Prompts executed in one batch (102 successful, 1 manual retry)

Zappy extension bulk generation result from workflow example

Timeline Editing: Sync Scenes to Audio Timestamps

Import the voiceover file and all scene images into your video editor. Drag images to the timeline in sequence. Use the filename timestamp (e.g., 'scene starts at 4 seconds') to position each image precisely. Move the playhead to the exact timestamp, trim the image from the right, and repeat for every scene. This creates the rhythm that drives viewer retention and is the core reason the stick-man channel succeeded.

Video editing workflow: timestamp-based scene synchronization

Claude Generates Viral Metadata in Seconds

Return to Claude AI and type 'Do you want to generate metadata?' Claude automatically generates a viral title, viral description, and viral tags optimized for your specific video. This metadata is ready to copy and paste directly into YouTube.

Generate Five High-CTR Thumbnails with One Click

Ask Claude to generate five high-CTR thumbnail prompts for your video (one line break between each) in a copyable code block. Copy all five prompts, open the Zappy extension, paste the prompts, set download folder, and hit generate. Five thumbnail options are created instantly. Pick the strongest one for upload.

5

High-CTR thumbnail options generated per video

Zappy extension bulk generation for thumbnail testing

End-to-End Free Workflow: Blank Page to Published Video

The complete workflow uses only free tools: Claude AI (script + prompts + metadata), 11Labs (voiceover), Faziscribe (timestamps), Google Imagen (images), and Zappy extension (bulk generation). No paid subscriptions, no design skills, no team required. From blank page to YouTube upload takes hours, not days.