AI video production toolkit — Dreamina, Remotion, ffmpeg, MiniMax TTS, powered by Claude Code.
A collection of battle-tested skills and workflows for producing AI-generated videos using Claude Code as the orchestrator. Covers the full pipeline from script writing to final render.
| Tool | Role |
|---|---|
| Dreamina | AI image & video generation (text-to-video, image-to-video) |
| Remotion | Programmatic video composition in React/TypeScript |
| ffmpeg | Video encoding, concatenation, subtitle rendering |
| MiniMax TTS | AI voice synthesis for narration |
| Claude Code | Orchestration — script, generate, compose, render |
Script → AI Image/Video Generation → Voice Synthesis → Composition → Render → QA
(Dreamina) (MiniMax TTS) (Remotion) (ffmpeg)
- Always verify renders — AI self-scoring is unreliable. Use
ffprobefor frame-level validation - CJK font rendering — Specify font paths explicitly in
drawtextfilters to avoid tofu (□□□) - Batch generation — Submit small batches to Dreamina, verify quality before scaling
- Credit management — Check
dreamina user_creditbefore expensive generation runs
- claude-jimeng-skills — Dreamina CLI skill for Claude Code
- evoskills — Self-evolving skill system
MIT