script to video, screenplay to video, shot list to video — turn a script into a storyboarded, shot-by-shot video: break down scenes, generate shots, assemble, validate. Use when the user gives a script or scene breakdown and wants a video.
设计与多媒体
video-to-screenplay
试用Turn video/audio transcripts into original screenplays. Faster-whisper → LLM story DNA → screenplay (.md + .fountain). Supports 6 LLM providers and CPU/GPU.
它能做什么
Turn video/audio transcripts into original screenplays. Faster-whisper → LLM story DNA → screenplay (.md + .fountain). Supports 6 LLM providers and CPU/GPU.
技能文档
video-to-screenplay
Turn video/audio/transcript files into original screenplays via a two-stage LLM pipeline.
When to use
- User wants to generate a screenplay, script, or story from a video/audio file
- User has a podcast, interview, or narration transcript and wants to develop it into a story
- User wants to extract story DNA (themes, characters, conflicts) from spoken content
- User needs Fountain-format output for import into screenwriting software
When NOT to use
- User just wants a transcript → use
transcribe_to_srt.py - User wants to summarize a video → use
web_fetch+ LLM - User wants to generate video from script → opposite direction
Pipeline
- Transcribe (optional): audio/video → SRT via faster-whisper
- Extract: transcript → story DNA (JSON) via LLM
- Generate: story DNA → original screenplay (.md + .fountain)
Requirements
- faster-whisper (
pip install faster-whisper) - One LLM API key in
openclaw.json(deepseek, kimi, zhipu, longcat, google, or agnes)
Usage
# From existing SRT
python scripts/video_to_screenplay.py --srt input.srt --out-dir ./output
# From audio/video (auto-transcribe first)
python scripts/video_to_screenplay.py --audio input.mp3 --out-dir ./output
# Specify genre and length
python scripts/video_to_screenplay.py --srt input.srt --out-dir ./output --target-minutes 15 --genre "科幻"
# Use GPU for transcription
python scripts/video_to_screenplay.py --audio input.mp3 --out-dir ./output --device cuda --compute-type float16
# Use a different LLM provider
python scripts/video_to_screenplay.py --srt input.srt --out-dir ./output --provider kimi --model kimi-k3
# Extract story DNA only (no screenplay)
python scripts/video_to_screenplay.py --srt input.srt --out-dir ./output --extract-only
Output files
transcript.txt— plain text transcriptstory_dna.json— extracted story elements (themes, characters, conflicts, motifs)screenplay.md— human-readable screenplayscreenplay.fountain— Fountain format for screenwriting software import
Provider support
| Provider | Default Model | Notes |
|---|---|---|
| deepseek | deepseek-v4-flash | Default, fast |
| kimi | kimi-k3 | temperature auto-set to 1.0 |
| zhipu | glm-5.2 | 1M context |
| longcat | LongCat-2.0 | Reasoning model |
| gemini-2.0-flash | Multimodal | |
| agnes | agnes-2.0-flash | Free tier available |
相关技能
Extract a storyboard table with local video frames
Converts articles (docx/pdf/txt/md) to narrated videos with AI-generated visuals and TTS voiceover. Invoke when user asks to convert article/document to video, mentions 文章转视频, or needs text-to-video with narration.
Generate timecoded SRT subtitles from local video or audio files. Use when a user wants a local low-cost subtitle workflow, asks to transcribe local media in...
18 次安装1 星标
用 Visla API 把脚本、网页、PDF 或音频转成视频。
100 次安装
article to video, text to video, news to video, essay to video — turn a written article into a narrated explainer video: outline, storyboard, voiceover, build, validate. Use when the user pastes or gives an article and wants a video.