GitHub repo digital-human-koubo lets users create AI talking‑head videos from a single photo
オープンソースプロジェクトは、Claude互換エージェントスキルを使用し、5つの編集スタイルと48個のパッケージングコンポーネントで、リップシンクされた多言語動画を生成します。
The repository digital-human-koubo, authored by andyhuo520, produces AI‑driven talking‑head videos from one photo and one sentence.
It builds a fixed digital presenter stored in persona.json, composites the figure into any scene, and generates ≤15‑second lip‑synced clips via Grok Video 1.5 on Renoise MCP.
The pipeline includes automatic credit quoting, Whisper transcription for quality control, voice unification processing, and final editing in Remotion with five selectable styles.
Supported output includes English, Japanese, Korean, French, Spanish and Arabic, with Whisper back‑check accuracy reported between 94 % and 100 %.
The skill folder follows the open Agent Skills format and works with Claude Code, Codex, OpenClaw, WorkBuddy and other agents that can load skills, connect to a remote MCP server and run local commands.
あなたにとっての意味
For creators, the tool offers a turnkey way to produce professional‑looking livestream ads or tutorials without filming, lowering production cost to a few Renoise credits. For developers of free or local AI models, the repo demonstrates how to integrate large‑language agents with video generation pipelines, expanding the range of applications that can be built on open‑source foundations.
ソース
- andyhuo520/digital-human-koubo: AI digital-human talking-head videos from one photo: Renoise MCP generation + 5 editing styles + 48 creator-grade packaging components. Agent skill for Claude Code / Co, GitHub trending, 26 Oct 2026
上記のソースから自動的に生成し、当社のコードでチェックしました:すべての数値と名前が含まれ、文がソースをコピーしてはいけません。間違いを見つけましたか?報告する。
今日無料で使えるAIはどれですか?
すべての無料AIモデルをブラインドヒューマン投票で順位付けし、あなたのGPUで必要なメモリと速度を示します。