capcut-cli

capcut-cli

English | 中文

Disclaimer: This is an independent, community-maintained project. It is not affiliated with, sponsored by, or endorsed by CapCut, JianYing, or ByteDance Ltd. "CapCut" and "JianYing" (剪映) are trademarks of ByteDance Ltd. All product names, logos, and brands are the property of their respective owners and are used here only for identification (nominative) purposes.

An independent, community-maintained CLI for CapCut / JianYing draft files.

An independent CLI for CapCut / JianYing that any LLM agent can drive — zero dependencies, no server, both namespaces in one binary.

Every command uses a version-aware local draft store: JSON in, JSON out, no MCP server or HTTP daemon. On newer CapCut versions it detects and synchronizes every readable timeline target instead of assuming draft_content.json is the only source of truth. That gives any model (Claude, DeepSeek, GLM, Kimi) a deterministic pipeline boundary for inspection, building, subtitles, captions, translation, and long-form cuts.

Use it three ways:

• CLI — npm install -g capcut-cli, then capcut <command> <project>
• Library — import { loadDraft, lintDraft, saveDraft } from "capcut-cli" (typed, zero-dep)
• Queue runner — capcut serve reads JSONL jobs from stdin and drops into n8n / Make / Coze

Run capcut doctor first to verify Node, whisper, ffprobe/FFmpeg capabilities, and draft directories.

See the changelog and releases for what's new in each version.

Current release: v0.11.3. The v0.11 line adds CapCut 8.7 multi-file storage, atomic/conflict-safe writes, transactional batch, command-schema v2, compile v2, karaoke captions, full media probing, multi-track proxy composition, and a concurrent retrying automation runner. v0.11.2 fixed Windows drive-letter paths across init, compile, serve, ESM consumers, and exact-byte restore; v0.11.3 synchronizes the English and Chinese documentation. See the v0.11.3 release.

Workflow

How capcut-cli fits into a typical viral-shorts pipeline. Steps 2 and 3 are LLM-driven (any model that returns JSON); steps 1, 4, and 5 are deterministic CLI calls. Step 6 stays human — every short-video platform forbids automated upload, so the publish click is yours.

flowchart LR
    A[Long video<br/>or CapCut project] --> B[capcut cut<br/>→ 60s candidate]
    B --> C[Claude / DeepSeek<br/>/ GLM / Kimi<br/>→ hook + script JSON]
    C --> D[capcut-cli<br/>add-text · add-audio<br/>apply-template]
    D --> E[CapCut / JianYing<br/>review + render MP4]
    E --> F[Publish<br/>YouTube Shorts · Reels · TikTok]

Comparison

How capcut-cli differs from the other CapCut / JianYing tooling: