Lumen
by Mas-inx
README.md
<div align="center">
<img src="docs/assets/logo.png" width="104" height="104" alt="Lumen logo" />
# Lumen — the open-source AI video editor
**Edit video with an AI that can see, hear and cut.**
<sub>Vibecoded with <b>Claude Opus 5.5</b> by <a href="https://github.com/Mas-inx"><b>Mas-Inx</b></a> at <a href="https://async.tech/"><b>AsyncTech</b></a></sub>
Lumen is a desktop video editor where the AI isn't a chatbot bolted onto a timeline. It watches your edit frame by frame, listens to the mix, reads what's said — and makes real, undoable edits with the same tools you use. Bring your own brain: **Claude Code**, **Codex**, or a model with your **OpenAI, Anthropic, Gemini, OpenRouter** or local **Ollama / LM Studio** key.
[](https://github.com/Mas-inx/lumen-ai-video-editor/releases/latest)
[](#build-from-source)
[](https://github.com/Mas-inx/lumen-ai-video-editor/releases/latest)
[](https://github.com/Mas-inx/lumen-ai-video-editor/releases)
[](https://github.com/Mas-inx/lumen-ai-video-editor/actions/workflows/ci.yml)
[](#credits)
<img src="docs/screenshots/editor.jpg" alt="The Lumen AI video editor: a chrome 3D title over a Blender-rendered background in the preview, the media library of generated clips, a timeline with captions, titles, overlays, three backgrounds, narration and music, and the Copilot panel showing its steps" />
</div>
## Why Lumen
- **An AI that edits, not just chats.** Every edit in Lumen is a typed command, and the AI gets all of them — plus tools to look and listen. More than 100 tools in all, each change a normal undo step. [See them all →](docs/AI-TOOLS.md)
- **It sees what you'll export.** When the AI checks a frame, it's rendered exactly as the export will be: every track, title, effect and transition. It can watch the whole edit as a contact sheet, look inside footage, and hear loudness, silences and speech.
- **Bring your own brain.** Use the Claude Code or Codex you already pay for, an API key from any major provider, or a local model. Lumen is also an **MCP server**, so any agent can drive it.
- **Local-first.** Media is edited in place and never uploaded. Speech-to-text runs on your GPU with Whisper. No account, no watermark, no telemetry.
- **A real editor underneath.** A magnetic timeline, keyframes, 3D titles and transitions, colour looks, frame-exact export with hardware encoding — so you can finish by hand whatever the AI starts.
It's an open-source alternative to editors like CapCut, Filmora and Descript — built so an AI can do anything you can.
<div align="center">
<img src="docs/screenshots/teaser.webp" width="720" alt="An animated teaser made entirely in Lumen: a chrome 3D title spinning in over Blender-rendered glossy shapes, a caption, and a 3D cube transition into the next shot" />
<br />
<sub>Made in Lumen, frame for frame: Blender backgrounds and a chrome 3D title, HyperFrames motion graphics, Whisper captions, and a 3D cube transition the Copilot added and checked.</sub>
</div>
## Contents
- [Features](#features) — [editing](#a-real-editor) · [Copilot](#the-copilot-sees-hears-and-edits) · [smart edits](#smart-edits) · [generation](#generate-3d-motion-graphics-images-video-and-voice) · [export](#export) · [integrations](#connect-anything)
- [Download](#download) · [Build from source](#build-from-source) · [Documentation](#documentation) · [Roadmap](#roadmap) · [Contributing](#contributing) · [Credits](#credits)
## Features
### A real editor
<img src="docs/screenshots/timeline.jpg" alt="Lumen's multi-track timeline: captions, a closing 3D title, overlay renders (3D title, lower third, kinetic type), three Blender backgrounds on the magnetic main track, narration and a ducked music bed, with markers on the ruler" />
- **Magnetic main track** that keeps your story edge to edge, with free tracks above and below for B-roll, overlays, titles, captions, voice, music and sound effects.
- Split, trim, ripple, slide into gaps, snapping, markers, speed and reverse, and **keyframes** on every property:
- 28 easing curves per keyframe, or your own cubic-bezier;
- one-click **motion presets** baked into keyframes: Ken Burns, push-ins, pans, pops, slides, and a punch that lands on the beat;
- keyframes you drag right on the timeline.
- **Render previews** like Premiere and Filmora:
- a bar under the ruler shows what may stutter and what's rendered;
- `Enter` renders in to out, or Lumen renders heavy stretches while you're idle;
- render at Full, 1/2, 1/4 or 1/8 resolution, sharpened back up on playback.
- **Smooth on modest computers.** Nothing redraws while idle, filmstrips and waveforms are painted once and reused, and the preview drops resolution when it has to.
- **Pro editing tools:** roll, slip and slide; ripple trim and insert or overwrite on any track; in and out points to play, lift, extract or export a stretch; copy, cut, paste and *paste attributes*; crop with rounded corners; freeze frames and **speed ramps**; detached sound for J- and L-cuts; groups; and solo and reorder for tracks.
- **2D and 3D transitions** (cube, flip, door, page curl, shatter…) placed before, on or after a cut, effects, **colour looks and grading**, and extruded **3D titles** rendered live.
- **Real colour grading** on the GPU: tone curves, lift / gamma / gain wheels, HSL by colour range, imported **.cube LUTs**, and **scopes** (waveform, RGB parade, vectorscope, histogram) beside the program monitor.
- **Compositing:** chroma and luma keys with an eyedropper and spill suppression; feathered rectangle and ellipse **masks** you drag on the canvas (on adjustment layers too, to grade just part of the frame); outlined titles in **any font** — imported, or picked from those installed on your computer, and embedded in the project.
- **Text behind subject:** on-device AI cuts the person (or, with a GPU, any subject) out of a shot. Slip a title between them and the background, remove the background, or remove the person. One click, one undo step, frame-exact on export.
- **Sound that drives the picture:** a Web Audio mix with per-clip volume, fades and keyframes, *Studio voice* clean-up, on-device noise reduction and 15 built-in sound effects.
- **A real mixer:** a channel per track and a master bus, each with a fader, pan, a draggable EQ curve, a compressor and a limiter, with live meters. Transitions crossfade the sound too, and exports can be **loudness-normalized** to −14 LUFS for YouTube and Spotify, −16 for Apple, or broadcast's −23 and −24.
- Import **MP4, MOV, WebM, MKV, MP3, WAV, FLAC, PNG, JPEG, WebP, GIF** and more — by dialog, folder or drag and drop. Files stay where they are; missing ones can be relinked. Select many at once (`Ctrl`/`Shift`-click or drag a box) to add, transcribe or remove them together.
- `.lumen` **project files** that survive moving folders between drives, crash recovery, and undo for everything. `Ctrl+K` reaches every command.
- **Built for long, heavy projects:** **proxies** for 4K footage (made automatically in the background; exports always use the originals), hour-long recordings that **stream** instead of filling memory, **version history** (every save, plus a snapshot every ten minutes while you edit, restorable any time) and **Collect project** to gather a project and all its media into one folder.
- **Multiple timelines and nesting:** as many timelines per project as you like — another cut, or a vertical copy reframed in one click — and **nest** any clips into a timeline of their own to grade, speed up or cut them as one piece (double-click to step inside; picture and sound both play through).
- **Multicam:** pick the camera recordings and Lumen **lines them up by their sound** (to about a millisecond), then press `1`–`9` while it plays to cut between angles, with every camera live in a strip under the preview.
- **Reads the footage:** **scene detection** splits or marks a clip at every shot change; **stabilization** smooths away camera shake with just enough zoom; **motion tracking** pins a title, sticker or mask to something moving — including a blur that follows a face.
### The Copilot sees, hears and edits
<table>
<tr>
<td width="46%"><img src="docs/screenshots/copilot.jpg" alt="The Lumen Copilot reading the project, watching a contact sheet, applying a batch of edits and checking a frame it rendered" /></td>
<td>
Ask in plain words — or talk to it. The Copilot plans, uses Lumen's tools, and checks its own work:
| You ask | It… |
| --- | --- |
| “Remove the long pauses and add captions” | transcribes on your GPU, cuts every silence across the whole timeline (captions and B-roll stay in sync), lays timed captions |
| “Make the intro punchier” | watches the edit, adds a 3D transition and a glow in one undo step, renders the frame to check it |
| “Is the music too loud under my voice?” | measures loudness per track in LUFS and ducks the music under speech |
| “Make a vertical version for Reels” | reframes the canvas and every layer to 9:16 |
| “Find where I talk about the mountains” | searches inside transcripts and jumps there |
| “Put the title behind me” | cuts you out of the shot on your computer and slips the title between you and the background |
| “Cut the B-roll to the music” | finds the beat and lays the clips out on it |
You see every step, including the frames it looked at, and **Undo all** reverts a whole answer.
</td>
</tr>
</table>
What the AI sees when it watches your edit — one call to `get_contact_sheet`:
<img src="docs/screenshots/contact-sheet.jpg" alt="A contact sheet the Lumen AI generated of the whole edit: labelled frames of 3D backgrounds, the chrome title, the lower third and captions" />
**Brains:** your own **Claude Code** or **Codex** (Lumen runs them headless, with only Lumen's tools), or a model with your key: **Anthropic, OpenAI, Google Gemini, OpenRouter, OpenCode Zen, OpenCode Go, Ollama, LM Studio** or any OpenAI-compatible endpoint. Pick the model and the **effort** — how hard it thinks, from Low to Max (Ultra on Codex) — for every one of them, Claude Code's Fable, Opus, Sonnet and Haiku and Codex's own model list included. Keys are encrypted by Windows (DPAPI) and never leave the app's main process.
- **Chats per project.** Start new chats, switch between them, and pick up where you left off after a restart.
- **Nothing happens out of sight.** The chat shows the model's thinking, each step while it is written, run and finished, and a line with a clock for what it is doing right now. Open any step to see what was sent and what came back.
- **Queue and stop.** Send more while it works and the messages wait their turn. Stop ends the reply at once, running exports and cut-outs included.
- **Attach files:** paste, drop or pick pictures, PDFs, Word documents and text files.
- **Prompt caching** on API keys: with Anthropic, each step re-reads the conversation from cache at about a tenth of the price, and every reply shows what it used.
- **Copy** any message, code block, or everything a reply did.
- **The web:** it can search, read pages, and look at a page as a screenshot, and links in its replies show previews.
- **Video links:** paste a YouTube or Vimeo link, a video post on X or a video file, and it watches it: frames across the whole video, the transcript from its captions, and the chapters. Nothing is imported.
- **Skills:** craft guides it loads before the work, for short-form editing, motion design, grading, sound, Blender and generation prompts. Add your own in the SKILL.md format. Claude Code, Codex and any agent on Lumen's MCP server use them too.
### Smart edits
One-click suggestions on the timeline, and tools for every AI:
- **Transcribe** on the device with Whisper (GPU via WebGPU, or CPU), or with OpenAI or ElevenLabs.
- **Remove pauses** — the whole timeline closes up around each cut, so captions, B-roll and markers stay in sync while music plays on; with several speakers it only cuts where everyone is quiet.
- **Captions** from the transcript, on their own track, in four styles: *Clean*, *Bold*, *Word by word* (key words in an accent colour) and *Boxed*. You can also import **SRT / WebVTT** files (drop them on the timeline), and export the captions as SRT or WebVTT, on their own or next to the video.
- **Find the beat** of a song: the tempo, beats and bars show on its waveform, cuts snap to them, and selected clips can be laid out on the beat.
- **Duck music** under speech with volume keyframes.
- **Reframe** between landscape, vertical, square and portrait.
- **Hear the mix** — loudness in LUFS against the −14 streaming target, peaks, clipping and silences.
### Generate 3D, motion graphics, images, video and voice
<img src="docs/screenshots/generate.jpg" alt="Lumen's Generate panel with motion graphic templates — lower third, title card, kinetic type and stat counter" />
- **Blender** — real extruded 3D titles, looping abstract backgrounds and AI-written scenes, rendered in the background and dropped into your media.
- **HyperFrames** motion graphics, rendered inside Lumen and pixel-exact at any display scaling. Choose a lower third, title card, kinetic type, stat counter, pull quote or chapter marker, or any HTML + GSAP composition. They're designed to look made, not generated: considered type, easing that differs per element, and holds long enough to read.
- **Images and video** from your OpenAI key (GPT Image) or Gemini key (Gemini image models, Veo video), with model lists fetched live — new models show up without an update.
- **Voiceovers, sound effects and music** with ElevenLabs.
- Anything from an **MCP server** — Higgsfield, Runway, Replicate, fal.ai or your own.
- **GTA V cinematics** from a FiveM game with **GS Cinematic Studio**: the Copilot stages the scene, directs the cameras, renders it in the game and drops the clip on your timeline.
- **Frame ingest:** another app on your computer sends finished frames straight to Lumen, which encodes each shot once into a full-colour (4:4:4) master, visually lossless or bit-exact. [How it works →](docs/INGEST.md)
### Effects, transitions and looks
<img src="docs/screenshots/transitions.jpg" width="49.5%" alt="Lumen's transitions library with previews of 3D transitions — cube, flip, doorway, swing in, page curl, warp zoom — and a cube transition mid-turn in the preview" /> <img src="docs/screenshots/looks.jpg" width="49.5%" alt="Lumen's colour looks — natural, cinematic, golden hour, nordic, vivid, faded film — previewed on the edit, with kinetic type in the preview" />
2D and 3D transitions, 15 effects (glow, grain, shake, RGB split, 3D tilt, curved screen…), one-click colour looks and full grading — each previewed on your own footage before you apply it.
### Export
<img src="docs/screenshots/export.jpg" alt="Lumen's export dialog with format, codec, resolution, frame rate, quality, hardware encoding and file size estimate" />
| Format | Codecs |
| --- | --- |
| MP4 | H.264, HEVC, AV1 |
| MOV | H.264, HEVC |
| WebM | VP9 (with **transparency**), AV1 |
| PNG | numbered frames, with **transparency** |
| GIF | animated, palette-optimised |
| WAV · M4A | audio only |
From 480p to 4K in your timeline's shape, 24–60 fps, with **hardware encoding** on Intel, NVIDIA and AMD graphics, a live file-size estimate and progress in the taskbar. Lumen tries the encoder with real frames first and falls back by itself: to the processor, then to another codec, even if the graphics card's encoder fails part-way through. *Graphics and encoding* in the Lumen menu shows what your computer can do. Export any timeline, burn captions in or save them as an **SRT / WebVTT** file beside the video, and line several exports up in the **render queue** to render one after another while you keep editing.
### Connect anything
<img src="docs/screenshots/integrations.jpg" alt="Lumen's integrations hub: AI models, Claude Code and Codex, Blender, HyperFrames, MCP servers and Lumen's own MCP server" />
- **Lumen is an MCP server.** Turn it on and any agent — Claude Code, Claude Desktop, your own — gets the same 100-plus tools, pictures included:
```bash
claude mcp add --transport http lumen http://127.0.0.1:47910/mcp --header "Authorization: Bearer <token from Lumen>"
```
- **Lumen is an MCP client.** Add any server by URL or command. Its tools become Copilot tools (Claude Code and Codex get them too), and generator tools also show up as models in the Generate panel.
## Download
**[Download the latest Lumen for Windows →](https://github.com/Mas-inx/lumen-ai-video-editor/releases/latest)**
Windows 10 or 11, 64-bit. The installer lets you install just for yourself (no admin needed) or for everyone, and associates `.lumen` project files.
Lumen keeps itself up to date: it checks the Releases page, downloads a new version in the background and installs it when you restart (or choose **Check for updates** in the Lumen menu).
> [!NOTE]
> The installer isn't code-signed yet, so Windows SmartScreen may say *“Windows protected your PC”*. Choose **More info → Run anyway**. Every release lists SHA-256 checksums.
Optional: [Blender](https://www.blender.org/download/) 4.2+ for 3D renders. Everything else is built in; the Whisper speech model downloads once (about 200 MB) the first time you transcribe on your device.
## Build from source
You need Windows 10/11 and [Node.js](https://nodejs.org/) 24.
```bash
git clone https://github.com/Mas-inx/lumen-ai-video-editor.git
cd lumen-ai-video-editor
npm install
npm run dev # the app, with hot reload
```
| Command | What it does |
| --- | --- |
| `npm run dev` | Electron app with hot reload |
| `npm run dev:web` | The UI in a browser tab (no desktop features) |
| `npm test` | Unit tests (Vitest) |
| `npm run typecheck` | TypeScript, renderer and main process |
| `npm run dist` | Windows installer in `release/` |
## Documentation
- [AI tools reference](docs/AI-TOOLS.md) — every tool the AI can use, with parameters
- [Frame ingest](docs/INGEST.md) — sending frames to Lumen from another app
- [Architecture](docs/ARCHITECTURE.md) — how the editor, media engine, AI and integrations fit together
- [Contributing](CONTRIBUTING.md) · [Changelog](CHANGELOG.md) · [Security](SECURITY.md) · [Third-party notices](THIRD-PARTY-NOTICES.md)
<details>
<summary><b>Keyboard shortcuts</b></summary>
| Keys | Action |
| --- | --- |
| `Space` · `J K L` | Play / pause · shuttle |
| `←` `→` · `↑` `↓` | Frame step · previous / next edit point |
| `S` · `Del` · `Shift+Del` | Split · delete · ripple delete |
| `V` · `B` · `N` · `M` | Select tool · blade · snapping · marker |
| `R` · `Y` · `U` | Roll · slip · slide tools |
| `I` · `O` · `X` · `Alt+X` | Mark in · out · around the selection · clear |
| `Shift+Space` · `;` · `'` | Play in to out · lift · extract |
| `Ctrl+C` · `Ctrl+X` · `Ctrl+V` · `Ctrl+Alt+V` | Copy · cut · paste · paste attributes |
| `Shift+F` · `Ctrl+L` · `Ctrl+G` | Freeze frame · detach audio · group |
| `Ctrl+Shift+D` | Crossfade with the clip before |
| `Ctrl+Alt+N` · `Ctrl+Alt+Enter` | Nest the selection · open a nested clip's timeline |
| `1` – `9` | Multicam: cut to that camera at the playhead |
| Drag + `Ctrl` · `Shift` | Insert · overwrite (on a trim handle, `Ctrl` ripples) |
| `=` `-` · `Shift+Z` | Zoom · zoom to fit |
| `Ctrl+Z` · `Ctrl+Shift+Z` | Undo · redo |
| `Ctrl+I` · `Ctrl+E` · `Ctrl+S` · `Ctrl+O` | Import · export · save · open |
| `Ctrl+Shift+E` | Render queue |
| `Ctrl+K` · `Ctrl+J` · `?` | Command palette · Copilot · all shortcuts |
</details>
## Roadmap
Ideas the project is heading towards — contributions welcome:
- [ ] macOS and Linux builds
- [ ] Code-signed installer
- [ ] More languages for the interface
## Contributing
Bug reports, ideas and pull requests are welcome — start with [CONTRIBUTING.md](CONTRIBUTING.md). Every edit in Lumen is a command in [`src/editor/commands.ts`](src/editor/commands.ts), and every AI tool lives in [`src/integrations/tools/`](src/integrations/tools), so adding a feature for people usually adds it for the AI too.
## Credits
Lumen was vibecoded with **[Claude Opus 5.5](https://www.anthropic.com/claude)** in [Claude Code](https://claude.com/claude-code) by **[Mas-Inx](https://github.com/Mas-inx)** at **[AsyncTech](https://async.tech/)**.
## License
[MIT](LICENSE) © 2026 Mas-Inx. Lumen builds on excellent open-source work — see [THIRD-PARTY-NOTICES.md](THIRD-PARTY-NOTICES.md).
This server cannot be deployed
Maintenance
ActivityMaintained
ResponsivenessResponsive