Where Are Video-Use Output Files Stored? The Complete Guide to Edit Directories

Video-use stores all generated output files in <videos_dir>/edit/, a directory created alongside your source footage that isolates renders, transcripts, and final videos from the codebase itself.

The browser-use/video-use repository follows a strict convention for file output: every session writes assets to a predictable subdirectory within your video source folder. This design keeps the tool's source code clean while giving users full ownership of their generated content.

The Standard Output Location: <videos_dir>/edit/

All video-use edits use a single, configurable root path resolved at runtime. The placeholder <videos_dir>/edit/ expands to an edit/ subfolder inside whatever directory contains your original video files.

According to SKILL.md, this is enshrined as a core design rule: "All session outputs in <videos_dir>/edit/" 【SKILL.md†L33】. The install guide reinforces this with the explicit statement that "all outputs land in <videos_dir>/edit/"【install.md†L146】.

What Gets Stored Where

Asset Type Relative Path Purpose
Final render edit/final.mp4 Fully stitched video after cuts, grades, and overlays
Slot renders edit/animations/slot_<id>/render.mp4 Individual renders for reuse or independent re-rendering
Transcripts edit/transcripts/<source-name>.json Whisper-generated SRT/JSON files per source clip
Packed transcripts edit/takes_packed.md Aggregated markdown file for easy review

The README.md confirms the most common user-facing path: your finished video appears as edit/final.mp4 next to your sources【README.md†L48】.

How to Locate and Access Output Files

Command Line Workflow


# Set your source footage directory

VIDEOS_DIR=/path/to/my_videos

# Run the skill (creates edit/ subdirectory automatically)

video-use run --videos-dir $VIDEOS_DIR

# Inspect generated outputs

ls "$VIDEOS_DIR/edit"

# → final.mp4  animations/  transcripts/

Rendering Individual Slots


# Render a specific slot (example: slot 3)

video-use render-slot --slot 3 --videos-dir $VIDEOS_DIR

# Output appears at predictable path

ls "$VIDEOS_DIR/edit/animations/slot_3/render.mp4"

Python Access Pattern

from pathlib import Path
import json

# Define paths relative to your video directory

videos_dir = Path("/path/to/my_videos")
edit_dir = videos_dir / "edit"

# Load a generated transcript

with open(edit_dir / "transcripts" / "clip1.json") as f:
    transcript = json.load(f)

print(transcript["segments"][0]["text"])

Implementation in Source Files

Several key files enforce this output convention:

This separation means you can safely delete or archive your <videos_dir>/edit/ folder without affecting the video-use installation, and conversely, updates to the tool never risk overwriting your work.

Summary

  • Primary location: <videos_dir>/edit/ — resolved at runtime to your source video directory
  • Key outputs: final.mp4, animations/slot_*/render.mp4, transcripts/*.json, takes_packed.md
  • Design principle: Generated assets live with source footage, never in the repository
  • Enforcement: Hardcoded defaults in transcribe_batch.py and documented rules in SKILL.md

Frequently Asked Questions

Can I change the output directory for video-use edits?

No — the <videos_dir>/edit/ path is fixed by design. The tool intentionally couples outputs to source locations for provenance and organization. If you need outputs elsewhere, symlink or move the entire <videos_dir> after processing.

What happens if the edit directory already exists?

Video-use appends to or overwrites files within edit/ as needed. Slot renders in animations/ are recreated on demand, while final.mp4 is replaced with each complete render. Transcripts are only regenerated if source files change or flags force reprocessing.

Why does video-use store outputs separately from the codebase?

This isolation prevents repository pollution and enables clean version control. As documented in SKILL.md, the convention ensures "only the user's <videos_dir> contains the edit artifacts" — critical for a tool meant to operate across arbitrary user projects.

Where do I find transcripts for my source clips?

Individual transcript files appear at edit/transcripts/<source-name>.json. For a consolidated view, check edit/takes_packed.md, which aggregates all transcript segments into a single markdown file for review or editing workflows.

Have a question about this repo?

These articles cover the highlights, but your codebase questions are specific. Give your agent direct access to the source. Share this with your agent to get started:

Share the following with your agent to get started:
curl -s "https://instagit.com/install.md"

Works with
Claude Codex Cursor VS Code OpenClaw Any MCP Client

Maintain an open-source project? Get it listed too →