Nano Banana 2 Lite (BudgetPixel)

Nano Banana 2 Lite is the budget end of Google's Gemini image stack, and the node's whole personality is in the flat 40-credit price. No size ladder to think about - it's fixed 1K output, one flat price per image, no matter what. That's the cheapest way to get a Nano Banana render in this pack short of the priciest tier of the other one. For high-volume work - generating a hundred concept variants, building a dataset, cranking out iteration after iteration - the flat pricing is what makes it the node you can stop thinking about.

It's a hosted API node like everything else in the BudgetPixel pack: no GPU, no weights, no model download. Your prompt goes to the API, the image renders server-side, and it comes back as a core IMAGE tensor in your graph. Closed model, no local option, filters baked in - all the usual Nano Banana ground rules apply, minus the resolution decisions.

The inputs

• prompt (required) - your text description.
• reference_images - up to fourteen, which is more than the non-Lite version allows. Multi-reference composition is where Lite actually gets interesting: you can feed it a small pile of images and it fuses them.
• aspect_ratio - fourteen options, 1:1 default.
• num_images - 1–4.

That's the whole interface. There's no size input because there's no choice - Lite is 1K, full stop. Output is images into Preview Image or Save Image.

How it slots against its siblings

In the pack's Nano Banana lineup, Lite is the quantity play. Nano Banana 2 proper starts at 50 credits for 0.5K and scales to 150 at 4K; Pro is 120–180 credits with heavier reasoning for "professional asset production." Lite is flat 40, fixed 1K, and up to 14 references - cheaper than the 0.5K tier of the main node. The trade is straightforward: you trade the size ladder and the Pro-level thinking for the best credits-per-image ratio Google offers through this pack.

If you're mostly doing reference-heavy composition at 1K, there's a real argument that Lite is all you need and the 2K/4K tiers are a tax you're not collecting value from. If you know you need high resolution for print or a hero asset, go up a tier instead.

Install and issues

Install via ComfyUI Manager (search "BudgetPixel") or:

cd ComfyUI/custom_nodes
git clone https://github.com/budgetpixel-ai/comfyui-budgetpixel.git
pip install -r comfyui-budgetpixel/requirements.txt

Restart, find it under Add Node → BudgetPixel → Image, and set your API key via BUDGETPIXEL_API_KEY or a budgetpixel.json file - no key widget, so workflows stay share-safe.

Standard pack caveats apply: a recent ComfyUI for core types, and hitting Stop abandons the local wait but the server-side render keeps running and still bills its flat 40 credits. And remember Lite is fixed at 1K - if you wire the output into an upscaler expecting to recover a 4K look, that's on you, not the node. The pack has a P-Image Upscale node for exactly that, if you need the pixels back.