DEDUPED Hot AI repos, turned into use cases Get the Monday 5

How to generate a voiceover with word timestamps from a script with Video Talkcraft

Use Fish Audio to produce narration audio and timing data from a JSON sentence list.

Short answer: Install the Python dependencies with pip install requests python-dotenv, and make sure ffmpeg is installed; then copy .env.example to .env and set FISH_AUDIO_API_KEY. Optionally set FISH_AUDIO_REFERENCE_ID to a voice ID from your voice library; create script.json with a sentences list, for example {"sentences": ["First sentence.", "Second sentence."]}. All steps.

  • Content
  • Intermediate
  • 30 minutes
  • From the README
  • Claude Code
  • Codex

Credentials. The workflow requires a FISH_AUDIO_API_KEY in a local .env file, so keep that file out of version control and out of shared folders.

Who it is forA creator with no recorded voice who wants a generated narration track that already has timing data for the video build.

The problemNarration audio and matching word timing usually come from separate tools, and the timing has to be re-aligned by hand.

The outcomeA WAV narration file, a timestamps JSON file, and a timing file ready for the Remotion project.

What you need first

  • Claude Code or Codex installed
  • Node 18 or newer and Python 3.10 or newer
  • ffmpeg installed
  • A narration script (text) and a finished voiceover audio file, or a Fish Audio API key if you want generated voice
  • Optional: a real person footage clip for on-camera shots

Steps

  1. Install the Python dependencies with pip install requests python-dotenv, and make sure ffmpeg is installed.
  2. Copy .env.example to .env and set FISH_AUDIO_API_KEY. Optionally set FISH_AUDIO_REFERENCE_ID to a voice ID from your voice library.
  3. Create script.json with a sentences list, for example {"sentences": ["First sentence.", "Second sentence."]}. (inferred, check the README)
  4. Run python3 scripts/tts_fishaudio.py script.json audio/full.wav audio/timestamps.json --timing-out remotion/src/timing.json for sentence-by-sentence mode.
  5. Listen to the audio before building, because the README says a match result does not confirm pronunciation or word-level accuracy.

Prompt for your agent

Use the video-talkcraft skill from https://github.com/Vincentwei1021/video-talkcraft to generate narration audio and word timestamps from <script.json> using Fish Audio. Expected output: audio/full.wav, audio/timestamps.json, and remotion/src/timing.json. Verify that the script and the aligned text match before continuing.

Caveat. The README says the default model is s2.1-pro-free, and its free quota and availability period are set by Fish Audio, not by this repo.

Use this in your AI agent

Ask your agent

Using the Deduped MCP, get the setup steps for Vincentwei1021/video-talkcraft and "Generate a voiceover with word timestamps from a script"

Add Deduped to Claude Code once

claude mcp add --transport http deduped https://dedupedai.com/mcp

Cursor, ChatGPT, others

Share: X LinkedIn Hacker News Email Copy link

Third-party repos we don't own. AI-written use cases can be wrong. Disclaimer