
Explore your creativity in one place
AI image generation, image editing, video generation, video editing and music generation - all in one studio.
- Local generation - freeCreate on your own graphics card. No accounts, no cloud, no usage limits.
- Remote generationNeed more power? Rent a GPU on RunPod or use the AI models on OpenRouter (paid services).
A helping hand at every step
- Stuck on the words?
- Story, lyrics or prompts not coming together? Work them out with your favourite LLM, right inside bitGen.
- A second pair of eyes
- Show your pictures to a vision model and get help describing them, prompting from them and making them better.
- Try another model in seconds
- Choose a style, then use the Basic Settings docker to swap the model, add LoRAs or change the sampler settings. No workflow editing needed.
One studio for the whole film
From one sentence to a whole film
A local language model writes the screenplay. The StoryBoard turns each scene into a shot, and the characters keep their faces and their own cloned voices from scene to scene.
- Films with dialogue, or music videos cut to the beats of a song
- MiniMax H3, LTX and Wan video models through ComfyUI
- Character swap, lip sync, VR footage and SeedVR2 upscaling

Cut it on a real timeline
Layers, keyframes, effects and adjustment layers on a preview canvas, with the tools you expect from an editor.
- Colour grading, masks and tracking, multicam and nesting
- Captions, karaoke with a bouncing ball, transitions and a track mixer
- Proxies for big footage; export when it is right

Paint, comp and design
An image editor with layers, masks, smart objects and adjustment layers - and the AI in it, for generating, refining and filling.
- Selections, painting, retouching, puppet warp and liquify
- Type with OpenType features, vector shapes and the pen
- CAD sketch layers with real dimensions and a parametric 3D body

Build the set
Sketcher is quick 3D modelling: draw, push and pull, then whole rooms and houses with shared walls, doors, windows, roofs and stairs.
- Floor plans, sections and printable sheets
- Lights and cameras, then render the room as a shot
- Blender, Daz Studio and Unreal Engine links for the heavy lifting

Score it and give it a voice
Songs, stems and sound effects made on your own card, and voices cloned from a few seconds of speech.
- Music with MiniMax Music3 and YuE2, and split stems
- Text to speech in your characters' voices, speech to text
- A sound effects docker that fills a scene with sound

Read it as sheet music
Any song becomes real sheet music: clefs, chords and the words under the notes. Press play and the notes and the words light up as they are sung.
- A score written from the lyrics, or transcribed from a recording
- Paper colours and textures, saved as a PDF to print
- Played with the song itself, or the written notes on their own

That is the short version. Every feature - 1098 of them - is listed with where to find it.
Your computer, your files
- Nothing leaves your machine
- The models run on your graphics card. Projects, pictures and films stay in folders on your own disk.
- No accounts
- No sign-up, no subscription and no credits to buy. The Community is the only part that goes online, and only if you join it.
- Fits the card you have
- Pick an edition for your card, from the editor alone to everything. 16 GB mode runs MiniMax H3 on a 16 GB card.
- In four languages
- The app, its guide and this site in English, 简体中文, 繁體中文 and 日本語. Change it from the menu bar at any time.
- Open source
- Every line is on GitHub under the MIT licence. Read it, change it, and send improvements back.
- Help from people
- Stuck? Ask in the Community's Help room, right inside the app, or in the Discussions on this site.
Don't have the hardware?
No big graphics card, or none at all? bitGen Studio can rent the power it needs and still run everything else on your own computer.
- RunPod: a GPU by the hour
- Pick a style and press Generate. bitGen makes a RunPod pod for it, installs ComfyUI, puts the style's models and node packs on it and renders there - your render starts by itself when the pod is ready. Stop the pod when you are done.
- OpenRouter: the language model
- The screenplay writer, ✨ Enhance and the prompt chat can run on OpenRouter instead of your card - one key, hundreds of models, many of them free.
- Only what you send goes out
- The editors, the timeline, your projects and files stay on your computer. Use either one for some jobs and your own card for the rest, or switch back to Local at any time.
Minimum system requirements
With a rented GPU or online services
RunPod renders, OpenRouter runs the language model - your computer runs the studio, the editors and the timeline.
| System | Windows 10 or 11 (64-bit), Linux or macOS |
|---|---|
| Processor | 64-bit, 4 cores (8 recommended) |
| Memory | 8 GB (16 GB recommended for the timeline and the image editor) |
| Graphics card | Any - built-in graphics run the editors |
| Disk | 20 GB free for the app and your projects - no models on your computer |
| Internet | Broadband. A fast upload helps when your own models or voice clips go to the pod |
| Accounts | RunPod (paid by the hour while a pod runs) and / or OpenRouter (many free models) |
On your own graphics card
Everything runs at home, with no accounts and nothing to pay for.
| System | Windows 10 or 11 (64-bit) or Linux. On macOS the editors run, but the AI video models need NVIDIA |
|---|---|
| Processor | 64-bit, 8 cores |
| Memory | 16 GB for pictures, 32 GB for video, 64 GB for character swap |
| Graphics card | NVIDIA with a current driver: 6-12 GB for the small AI, 16 GB for MiniMax H3 in 16 GB mode, 20 GB or more for everything |
| Disk | 100 GB free, on an SSD - ComfyUI, the models and your projects |
| Internet | Only to download the app and the models once |
Mix them: render on your card and send only what it cannot do to a pod. The hardware guide says what each card can do.
Get bitGen Studio
One small installer sets up everything: the app, ComfyUI, the node packs and the models for your card.
| System | Windows 10 or 11, Linux, macOS |
|---|---|
| Graphics card | NVIDIA with a current driver. 6-12 GB for the small AI, 16 GB for MiniMax H3 in 16 GB mode, 20 GB or more for everything. The editors need no AI at all. |
| No graphics card? | Render on RunPod and run the language model on OpenRouter. |
| Disk | 100 GB free for the full set of models |
| Price | Free |