Images

​

Understand Images

Vision is Notis’s image understanding capability that enables our AI to “see” and analyze the content of images you share. When you upload an image, the Vision tool processes it and can describe, analyze, or extract information from the visual content.

​

Features

• Understand image content (people, objects, scenes, text, charts, screenshots)
• Extract structured information (business cards, receipts, product labels, handwritten notes)
• Analyze visual elements (data visualizations, design elements, concepts)

​

Examples

“Extract information from this restaurant receipt, including the restaurant name, date, time, items purchased, and payment details.” “Digitize this business card information including name, title, company, contact details, and website.” “Analyze this quarterly sales chart and provide a detailed breakdown of the numbers along with growth percentages between quarters.” “What is this plant and how much should it be watered every week?”

​

Generate & Edit Images

Notis can both create new images from a prompt and edit images you’ve already shared. Under the hood it uses two state-of-the-art models and automatically picks the right one based on the controls you ask for, so you don’t have to think about it.

​

Models

Nano Banana Pro — Google’s gemini-3-pro-image-preview. Built for professional asset production, with advanced reasoning that follows complex instructions and renders high-fidelity text inside images. Strong choice for infographics, diagrams, posters, and any image that needs accurate typography or that combines multiple reference images. ChatGPT Image — OpenAI’s gpt-image-2. A strong all-rounder with transparent-background support, multi-output generation, and a high-fidelity editing mode that preserves faces, logos, and fine textures from reference images.

​

Shared features

• Generate images from voice or text instructions.
• Edit images you sent, that Notis created previously, or that live in your Notion databases.
• Save the result directly into a Notion media property in the same turn.

​

Nano Banana Pro capabilities

• Aspect ratios — 1:1, 2:3, 3:2, 3:4, 4:3, 4:5, 5:4, 9:16, 16:9, 21:9.
• Resolution tiers — 1K, 2K, or 4K for poster- and print-quality output.
• Multiple reference images — combine several inputs in a single edit (e.g., place a product on a new background, mix two style references, blend characters).
• Google Search grounding — optional real-world grounding so the model can render up-to-date facts (logos, charts, places, current events) instead of hallucinating them.
• High-fidelity text — reliably renders readable text and labels inside the image, ideal for infographics, slides, and UI mockups.

​

ChatGPT Image capabilities

• Sizes — 1024x1024 (square), 1536x1024 (landscape), 1024x1536 (portrait), or auto to let the model decide from the prompt.
• Quality — low (faster), medium, high (best quality), or auto.
• Transparent background — generate or edit assets as transparent PNGs, ready to drop onto any background.
• Multiple outputs — produce up to 4 variations in a single call.
• High input fidelity — when editing, opt into a high-fidelity mode that preserves details like faces, logos, and fine textures from the input images.

​

Examples

Generate new images: “Generate an image of a cat riding a bicycle in a cartoon style.” “Create a 16:9 4K infographic of this quarter’s sales numbers and save it in the media property.” “Make a transparent PNG of a blue rocket I can drop into a slide.” “Generate 3 variations of a minimalist logo for a coffee shop.” “Create a poster of the current weather in Tokyo.” (uses Google Search grounding for real-world data) Edit existing images: “Edit this image to add a rainbow in the background.” “Combine these two product photos onto the same beach background.” “Remove the text from this image and replace it with our company logo, keeping the logo crisp.” “Change the background color to blue while keeping the subject’s face exactly as it is.”