AGENT SKILL

export-nle

BETA (not yet tested in actual NLEs): Export edit.json to Final Cut Pro / DaVinci Resolve (FCPXML), Premiere Pro (FCP7 XML), and SRT captions. Use for requests to open in Premiere or Final Cut, export for Resolve, or get an SRT. Always report fields that cannot transfer in dropped[].

License 许可
MIT
最近更新
2026-10-22
Stars
★ 214
Forks
25

01 前言 Front Matter

name
export-nle
description
BETA (not yet tested in actual NLEs): Export edit.json to Final Cut Pro / DaVinci Resolve (FCPXML), Premiere Pro (FCP7 XML), and SRT captions. Use for requests to open in Premiere or Final Cut, export for Resolve, or get an SRT. Always report fields that cannot transfer in dropped[].

03 SKILL.md 正文

Export editing data to other NLEs (BETA)

Language: Respond in the user's language — conversation, questions, approval prompts, and the reports and records you write (including decision-log rows) follow the language the user writes in.

⚠ BETA — implemented, but not verified in actual applications This feature has been implemented, but its output has not yet been imported into actual Final Cut Pro, DaVinci Resolve, or Premiere Pro applications for verification. Reports of XML that fails to import or is misaligned are welcome as issues. Always tell users about this beta status when handing over results.

What this is

A one-way export demonstrating that AKARI Video's saved-data SSOT (edit.json) has no lock-in. It gives users who want to finish in a professional tool an exit path.

Format Target File
FCPXML 1.11 Final Cut Pro / DaVinci Resolve <project>.fcpxml
FCP7 XML (xmeml v5) Premiere Pro <project>.premiere.xml
SRT Caption sidecar shared by all NLEs <project>.srt (only when captions.json exists)

The reverse direction (bringing NLE edits back into edit.json) is unsupported. It is outside scope.

Do not read all of edit.json or captions.json; grep by ID and read only the matching lines (how to read edit.json). This skill does not write them. If another stage must write them, use Edit on the matching lines or the edit-store script API.

Hard rules

  1. Be deterministic: identical input → identical output. Do not mix LLM judgments, randomness, or current time into output.
  2. Zero external npm dependencies (call the ffprobe binary directly through media-bin only).
  3. Target only version 2 edit.json; grep by ID to read matching lines and do not modify it. Run akari migrate on older versions first.
  4. Never silently drop fields: enumerate every field that the exchange format cannot carry (ducking / master / LUT / chroma_key / direction, etc.) in export-report.json dropped[].
  5. Explicit beta status: Always state that actual NLE import has not been verified in the user report. Never present it as verified.

Procedure

  1. Run against the target project (the directory containing edit.json).

    node packages/export-nle/bin/export-nle.mjs <project-root|edit.json>
    

    The default destination is <project>/exports/nle/. To limit formats, use --format fcpxml (comma-separated values allowed); use --json for machine-readable output or --no-probe to avoid ffprobe.

    Input must be edit.json v2. Convert a v0/v1 project to v2 with akari migrate first.

  2. Check the exit code. 0 means export complete (warnings still count as success); 2 means an input or runtime error.

  3. Read exports/nle/export-report.json and include dropped[] (non-transferrable fields) and warnings[] (placeholder durations, approximations, etc.) in the user report.

  4. Media is written as references (absolute-path file URLs). Tell users they must reconnect media with the NLE's relink feature after moving or sharing the project.

What transfers and what does not (contract summary)

Canonical source: contract-2026-08-01-export-nle-beta.md

  • Transfers: tracks[].items[] media (interpret tracks[] array order as bottom-to-top z order; convert integer-frame at / duration to NLE time; preserve source in / out, speed, transform, and opacity; use stable item.id for the NLE clip name; carry blend into FCPXML and warn for xmeml), transition_out (dissolve directly, fade-black/white approximated by cross dissolve), pre-baked html / telop (alpha MOV, etc., as ordinary media clips), narration / sfx / bgm (place from the tracks[] audio lane or edit.audio.*, with gain; expand bgm loops and fades), and captions.json (as plain-text SRT).
  • Does not transfer (report in dropped[]): ducking, audio.master (loudnorm / denoise), output.look (LUT), chroma_key, and caption styling (karaoke treatment and background plate).

source.kind policy

source.kind NLE handling
media Export as a media clip resolved through sources[]. Use the item's integer frames for timeline time and second-based in / out for the source window.
html Export as a media clip if a baked asset is present in the internal representation. Current v2 html declarations lack baked, so normally omit it and put the reason and baking procedure in dropped[].
telop Export source.baked as a media clip if present. Otherwise omit it and record the reason and baking procedure in dropped[].
filter No interoperable clip representation exists. Omit it and record it in dropped[]; bake it into video if necessary.

Never disguise unbaked html / telop as media references to HTML paths or preset names. Export narration / sfx / bgm items in audio lanes from tracks[]. For documents without audio items in tracks[], continue to read edit.audio.narration / sfx / bgm. Record omitted audio items, including muted items, speech roles, and the second or later bgm item, with reasons in dropped[]. For tracks[] sfx, renumber NLE tracks from 0 in the order of audio lanes containing sfx, closing any gaps. For ducking, prefer edit.audio.bgm.ducking when present; otherwise read the exported bgm item. Continue to report it as audio.bgm.ducking in dropped[] (after migration it resides on the bgm item). Continue reading audio.master from top-level edit.audio.master and report it in dropped[]. A bgm item's at > 0 does not transfer because the exporter places it at 0 seconds; report it in dropped[] as tracks[…].items[…].at.

Out of scope

  • CapCut draft format (unofficial, increasingly encrypted, and liable to break with version updates; revisit through a lab/ spike if demand appears).
  • Reverse import from an NLE.
  • Automation of actual application import verification for generated XML (required to leave beta; design it after manual verification results accumulate).

04 同仓库更多 Skills

AkariLabs
AkariLabs/akari-video

verify-project

Verify an AKARI Video project with the L0 / L1 / L2 verification ladder required by an AKARI Video task contract on the current Theia stack. Read when acceptance criteria specify levels such as L0 or L0+L1, to determine exactly what to run and how at each level.(旧名: verify)

★ 214
AkariLabs
AkariLabs/akari-video

verify-edit

Verify edit.json and optional analysis.json / captions.json / media with a deterministic CLI, then finish QA with frame inspection and reporting after PASS. Use immediately after writing or modifying edit.json, before export, and after applying review feedback.(旧名: edit-lint)

★ 214
AkariLabs
AkariLabs/akari-video

setup-remote

Set up remote access to approval reports and previews from another device such as a phone, and send footage to the workspace. Use doctor to assess Tailscale, guide installation and login (human actions), expose the preview server (default 4567) and approval report helper over tailnet-only HTTPS with tailscale serve, connect Taildrop reception to workspace inbox/, and verify access from another device. Use for requests to view reports on a phone, approve remotely, send footage from a phone, or set up remote access. Public internet exposure (funnel) is outside the default scope.

★ 214
AkariLabs
AkariLabs/akari-video

setup-project

Set up an AKARI Video project headlessly. Recursively copy `templates/project-default/`, record the template version, initialize git only when safe, and generate an HTML creation report. No app launch is needed. Use to create a video project or fill in an existing folder as an AKARI Video project.(旧名: create-project)

★ 214
AkariLabs
AkariLabs/akari-video

setup-library

Use for first-time AKARI Video setup or when the current project needs more usable assets. This first-run skill checks ffmpeg / whisper-cli (and can install with approval), reads the bundled catalog with akari assets to propose a starter pack, obtains explicit human approval, then acquires, places, validates, and updates INDEX entries. It also covers the audio library (BGM and sound effects): bulk acquisition of first-party AKARI Sounds from GitHub Releases, a free-source candidate HTML list with drop-folder matching or assisted acquisition for missing families, and keep/drop decisions in the listening gallery. Use for requests such as "I want more BGM" or "add sound effects to the library".(旧名: setup-audio-library)

★ 214
AkariLabs
AkariLabs/akari-video

setup-connections

Set up and inspect AKARI Video connections: AI-service keys, which models are allowed, cost-approval policy, the tool and media-service state, and 3D settings. Use for first-time setup, adding a provider or key later, checking what is usable (「何が使える?」 'what can I use?'), setting a key such as fal (「fal の鍵を設定したい」), 3D settings (「3D の設定を使いたい」), connection checks, and gates before paid generation or external publication. Model selection belongs to the production skills (they read the ledger and preferences); this skill decides allowed models and the cost policy. Works with `.akari/connections.json` and a free, read-only doctor.(旧名: manage-connections)(旧名: connect)

★ 214