Img2Img
Transform photos with multiple AI models, no sign-up require
I started building Img2Img in mid-July because image-to-image AI felt more fragmented than it needed to be.
The same source photo and prompt can produce completely different results depending on the model. But comparing those results usually means switching between several tools, creating multiple accounts, and repeating the same setup each time.
I wanted one focused workflow:
1. Upload an existing photo or image
2. Describe what should change
3. Specify what should remain recognizable
4. Try different AI image models
5. Compare the results and choose the direction that works best
That became Img2Img.
A few product decisions I made early:
- The free model works without creating an account
- Multiple AI models are available in the same interface
- Source and result are shown together for easier comparison
- Paid usage is based on one-time credit packs instead of another subscription
- The product focuses on transforming existing images rather than generating everything from scratch
The most interesting thing I have learned so far is that model choice can matter as much as prompt quality. A model that works well for a portrait may not be the best choice for a product photo, illustration, interior, or background replacement. Making those differences easier to explore is becoming the main value of the product.
The product is still early, and I am currently trying to answer three questions:
1. Is comparing multiple models the strongest reason to use the product, or would most users prefer one recommended default?
2. How many controls should be exposed before the workflow starts to feel complicated?
3. Which use case should I focus on first: portraits, product images, marketing creatives, or general image editing?
If you regularly work with AI images, I would appreciate honest feedback—especially on where the workflow feels confusing or unnecessary.
You can try it here without signing up:
2 Comments
“No signup” is a strong hook.
The interesting part is whether users come back after that first use.
Are you seeing repeat usage yet?
What stood out to me is that you're questioning whether people actually want more model choice or simply more confidence in the result they'll get.
Those lead to very different products. Sometimes reducing the number of decisions creates more value than expanding the number of options.