Features
Everything Kitly tracks, so you don't have to.
One declarative catalog of pipelines across Image, Video, Audio, and 3D, and one consistent record of how every asset came to exist.
30+ pipelines, one shared catalog
Declared once, not hand-wired per pipeline — new ones show up everywhere a pipeline picker does.
- text-to-image
- image editing (the nano-banana family)
- background removal
- depth estimation
- upscaling
- image-to-3D
- Gaussian splats
- image-to-video
- video-to-video
- sound effects
- text-to-speech
The Kitly Asset Library
One entity, five channels, every variation kept.
Treat a creative entity — “the goblin warrior I'm designing” — as one container spanning Text, Image, Audio, 3D, and Video, each with its own prompt, presets, and references. Every channel keeps a gallery of variations you can rate, comment on, and chain into each other.
Full lineage
Every generated asset knows the operation, provider, endpoint, and settings that produced it, and its parent asset. Pick up any variation and keep going from there.
Background job queue
Generations run async with live status. A job that survives an unclean shutdown resumes cleanly on next launch — nothing silently drops.
A real file manager
Rename, duplicate, delete, reveal-in-Explorer, copy and paste, right-click context menus. The Viewer supports zoom, pan, screenshot, and clipboard — it's a workspace, not just a gallery.
Plain files, your disk
Every project lives entirely on your local filesystem, in plain files you can open outside the app. No proprietary format, no lock-in.
Motion capture & retargeting
Turn a video into animation on your own rig.
Record body or facial motion from any video and Kitly tracks it frame by frame. Bring the capture into Retarget mode, where an auto-mapped bone table lines it up against your character rig — adjust individual bones, smooth the result, and save the mapping to reuse on the next character.
The Assistant
A tool-calling chat that runs the pipelines for you.
Ask for what you want and the Assistant creates Kitly Assets and triggers generations on your behalf. Plan mode drafts starting prompts for a goal across channels. Conversations persist across restarts, scoped per-project or per-asset.
Bring your own key
Kitly doesn't bundle AI usage into its price.
Connect your own API keys and pay each provider directly, at their own rates — or skip keys entirely with models that run locally.
Fal.ai
The image, video, audio, and 3D pipelines in the catalog above.
Google AI Studio
Gemini models, for the Assistant and Plan mode.
OpenAI
For the Assistant and Plan mode.
OpenAI-compatible endpoints
Any provider or self-hosted server that speaks the same chat-completions API.
Local ONNX models
Upscaling, depth, and segmentation run on your own machine — no key, no internet, no per-generation cost.