My Blog

Tag: video assembly)

  • AI Tools: I Fully Automated Google Flow Using Claude Code (Full Walkthrough)

    Watch on YouTube: https://www.youtube.com/watch?v=JEla_mdlfxs

    Channel: Guney Topcu | AI Made Easy

    Overview

    The video demonstrates a fully automated workflow for generating Google Flow videos at no monetary cost by combining the open‑source Flow Studio tool, Claude Code, and the Model Context Protocol (MCP). It shows how creators can stay within Google Flow’s free daily allowance of 50 credits while producing narrated short films using only free APIs and local FFmpeg processing.

    What the Video Covers

    The presenter walks through the entire end‑to‑end process:

    • Installing and configuring Flow Studio via Claude Code.
    • Leveraging Google Flow’s 50‑credit free tier (15 credits per 10‑second video shot, zero credits for still images and character creation).
    • Obtaining AI voiceovers at zero cost by routing Flow Studio to Google AI Studio’s free Text‑to‑Speech API instead of Flow’s native TTS (which costs 7 credits per clip).
    • Assembling the final video locally with FFmpeg, incurring no credit usage.
    • Following the five‑step pipeline: CAST → FRAMES → SHOTS → VOICE → CUT.
    • Using Exact Join for background continuity and Character Lock for location jumps.
    • Accessing a local dashboard at 127.0.0.1 with seven tabs, each corresponding to a workflow stage, or issuing plain‑English commands through Claude Code.

    Step‑by‑Step Walkthrough

    1. Setup
      • Install Node.js v22+ if not already present.
      • Ensure Google Chrome and FFmpeg are available (the installer can add FFmpeg automatically).
      • Run the Flow Studio installation command from the GitHub repo (github.com/GTAI-1/flow-mcp).
      • During setup, a dedicated Chrome profile opens; log in to Google Flow.
      • Create a free API key in Google AI Studio (aistudio.google.com) and paste it into the VOICE tab of Flow Studio.
    2. Verify Connection
      • Ask Claude Code: “Check Flow and tell me my credit balance.” The response should show the current free‑credit count (starting at 50 per day).
    3. CAST – Character Creation
      • In the CAST tab (or via Claude), describe the desired character in plain English.
      • Character creation and edits cost zero credits, allowing unlimited iterations.
    4. FRAMES – Keyframe Stills
      • Generate still images that will serve as keyframes for each shot.
      • Stills are produced with the Nano Banana 2 model and incur zero credits.
      • Review the outputs; request adjustments until the visual composition matches the vision.
    5. SHOTS – Video Clips
      • Convert each approved keyframe into a 10‑second video shot.
      • Each shot consumes 15 credits (Omni 1.1 Flash model).
      • Choose Exact Join when consecutive shots share the same background to maintain continuity; use Character Lock when the scene changes location.
      • Enable the credit safety guard (e.g., set a personal limit of 25 credits) to prevent accidental overruns; the system will block a shot that exceeds the limit and display a message like “OVER YOUR LIMIT: Not generated. 60 > 25 credits.”
    6. VOICE – Narration
      • In the VOICE tab, input the script for narration.
      • Route the request to Google AI Studio’s free TTS API, resulting in zero credit usage.
      • If the native Flow TTS were used, each clip would cost 7 credits.
    7. CUT – Final Assembly
      • Download the generated video clips and audio files.
      • Use FFmpeg locally to concatenate clips, overlay the voiceover, and add any desired transitions or effects.
      • This step incurs no Google Flow credits.
      • Export the final video in the desired format.

    Tools, Numbers, and Limits

    • Google Flow free tier: 50 credits per day.
    • Cost per 10‑second video shot: 15 credits.
    • Still images and character creation/edit: 0 credits.
    • Native Flow TTS voiceover: 7 credits per clip.
    • Voiceover via Google AI Studio API: 0 credits.
    • Restyle edit (modifying an existing clip): 40 credits – significantly higher than generating a new 10‑second clip.
    • Flow Studio provides 39 presets (30 camera moves, 9 visual effects).
    • Required prerequisites: Node.js v22+, Google Chrome, FFmpeg, a Google Flow login, and a free Google AI Studio API key.
    • Local dashboard accessible at 127.0.0.1 with seven tabs matching the workflow stages.

    Caveats and Hype Check

    “I made both of these clips for free without sitting inside Google Flow writing prompt after prompt.”

    The workflow is cost‑free only as long as the user remains within Google Flow’s daily 50‑credit allowance. Heavy usage — particularly frequent restyle edits at 40 credits each — can quickly exhaust the free tier, after which additional credits would need to be purchased or the user would need to wait for the next day’s allocation. The video does not make unverified revenue claims; it simply contrasts the zero‑cost approach with typical commercial Chrome extensions that charge around $30 per month.

    Concrete Next Actions

    • Install Node.js v22+ if not already installed.
    • Run the Flow Studio setup command from the GitHub repository.
    • Confirm that Chrome and FFmpeg are available (the installer can add FFmpeg).
    • Log in to Google Flow via the Chrome profile launched during setup.
    • Obtain a free API key from Google AI Studio and paste it into Flow Studio’s VOICE tab.
    • Test the connection by asking Claude Code to report your current credit balance.
    • Begin character creation in the CAST tab using natural‑language descriptions.
    • Generate and refine keyframe stills in the FRAMES tab.
    • Produce video shots in the SHOTS tab, selecting Exact Join or Character Lock as appropriate.
    • Create the voiceover in the VOICE tab, ensuring the Google AI Studio API key is active.
    • Assemble the final video locally with FFmpeg in the CUT tab.