Ember: Local AI Image Studio
A private, on-device image studio for generation, editing, projects, and video.
스크린샷
기능
Local Generation
Create images from text prompts entirely on your own hardware.
Word-Based Editing
Use up to three reference images and describe changes in plain language without masks, layers, or brush work.
Project Queues
Plan a shot list, queue an entire project, and retry failed shots with one click.
Narrated Video Export
Combine images and narration into an MP4 with automatically aligned timing, selectable voices, languages, and frame rates.
Organized Library
Search and filter images by their prompts, sizes, seeds, and models, then export individual pictures or selections.
Model Management
Download starter models or register model files already stored on disk without copying, moving, or altering them.
소개
PLEASE NOTE: This app runs image generation entirely on your own machine, and a dedicated graphics card is required for usable speed. On a computer without one, generating a single image can take many minutes. Please check your hardware before buying. Ember turns your computer into a complete image studio. Write a prompt, get a picture. Point at a picture you already have and describe the change you want. Everything happens on your own hardware, so nothing you type and nothing you create is ever uploaded, queued on someone else's server, or counted against a monthly credit balance. There is no account to create and no subscription to renew. Install it, add a model, and start working. PRIVATE BY DESIGN Your prompts, your reference images and your finished pictures stay on your machine. There is no telemetry and no sign-in. Work on confidential product shots, unreleased designs or personal photographs without wondering where they end up. GENERATE FROM A DESCRIPTION Describe what you want and Ember renders it. Choose a shape that fits where the picture is going — widescreen, vertical, square, or classic photo proportions. Reuse a seed to keep a look consistent across a set, or leave it random to explore. EDIT WITH WORDS INSTEAD OF TOOLS Bring in up to three reference images and describe the change in plain language. Restyle a photograph, change a background, combine elements from separate pictures, or carry one subject into a completely new scene. No masks, no layers, no brush work. BUILD A WHOLE SET IN ONE RUN The Story Planner screen turns a single topic into a production plan. Ember writes a prompt you paste into any AI chat, and the plan that comes back imports straight into a project. Each project keeps its own shot list, so several productions can sit side by side without getting mixed up. Queue the whole project and let it run. Ember works through the list one shot at a time and keeps going while the window sits in the background. Anything that fails can be retried in a single click rather than hunted down shot by shot. TURN A PROJECT INTO A NARRATED VIDEO Every shot carries the narration that is spoken over it. Ember records that narration in a natural voice, holds each image on screen for exactly as long as its own line takes to say, and exports the result as a finished MP4. Picture and voice line up without any manual timing. Pick the language and the voice, choose a frame rate, preview the result, and save it wherever you like. A LIBRARY THAT STAYS ORGANISED Every picture you make is kept with the prompt, size, seed and model that produced it. Search past prompts, filter by how a picture was made, and reuse any result as the starting point for the next one. Export a single picture or a whole selection to any folder in one step. YOUR MODELS, YOUR CHOICE Download a compact starter set from inside the app, or register model files you already keep on disk. Registered files are never copied, moved or altered — Ember simply reads them where they are. Downloads resume where they left off if the connection drops. Ember picks sensible generation settings for whichever model is active, so there are no sampler dials to learn before the first good result. WHAT YOU NEED The inference engine and the video encoder are both included in the app — there is nothing to install separately and nothing is downloaded to make the app work. You only choose a model: download a starter set from inside the app, or point Ember at model files you already have. Generation runs on your graphics card when you have one with enough memory. Without a dedicated card the work falls back to the processor, where a single image can take many minutes. The app shows which device it is using and how much memory is free, and tells you plainly when your hardware is not enough rather than failing silently. Narration voices are fetched over the internet when you export a video; everything else works offline.
자주 묻는 질문
Does Ember upload prompts or images?
No. Prompts, reference images, and finished pictures stay on your machine, and there is no telemetry or sign-in.
Do I need a dedicated graphics card?
A dedicated graphics card is required for usable speed. Without one, generation falls back to the processor and a single image can take many minutes.
Can I edit images without manual tools?
Yes. Bring in up to three reference images and describe the changes in plain language, without masks, layers, or brush work.
Can Ember create narrated videos?
Yes. It records narration, matches each image's display time to its spoken line, and exports the result as an MP4.
Can I use my own model files?
Yes. You can register model files already stored on disk, and Ember reads them without copying, moving, or altering them.
개인정보 보호 및 요구 사항
- 개인정보 보호
- Prompts, reference images, and finished pictures stay on your machine; there is no telemetry, sign-in, or cloud upload of your work.
- 권한
- Internet access is used to download starter models and fetch narration voices for video export.
- 언어
- Chinese (Simplified), Chinese (Traditional), English (United States), French, German, Italian, Japanese, Korean, Spanish