01
An image must be reproducible
Every generation hands back the seed, steps and guidance that made it. If you cannot make the same picture twice, you do not have a tool — you have a slot machine.
Home / About
The companyWe build the model.
Not a wrapper.
Paintune trains and evaluates its own diffusion models. That is the whole reason we can fix the things people actually complain about.
Why we exist
Most people can describe the picture they need
They just cannot make it. The gap between those two things is where projects stall, budgets go, and good ideas quietly get dropped.
Closing that gap is the entire company. Not a chat interface bolted onto somebody else’s checkpoint — our own models, tuned against the failures that make people throw a generation away.
How it works
What we hold to
Three rules we will not trade away
02
Consistency beats novelty
One striking image is easy. Forty that look like they belong together is what real work needs, and it is what most models are worst at.
03
Provenance is not optional
Generated images should say so. We write content credentials into the file rather than leaving it to whoever downloads it to work out where it came from.
Roadmap
Where we are, and what is next
The core is shipped. What follows is depth of control and reach into other products.
Phase 01Shipped
Foundation
- Text-to-imageCore generation, live and in production.
- Reproducible outputSettings returned with every image.
- Print resolutionUpscaling to full delivery size.
Phase 02In progress
Control
- Style LockFix a look once and hold it across a set.
- InpaintingChange one region, keep the rest.
- BatchesOne prompt, many seeds and ratios.
Phase 03Planned
Platform
- Public API and SDKsGeneration inside your own product.
- WebhooksBuilt for pipelines, not people.
- Scaled capacityMore throughput, lower latency.
Come and break it
We learn most from the prompt that fails. Tell us the image you cannot get.