Generate images

Introduction

remio's image generation is a visual thinking assistant, not just a tool for creating art . It transforms complex text descriptions into diagrams (like flowcharts and architecture diagrams) or illustrations, designed to help you understand, organize, and retain knowledge more efficiently.

• Visualize Logic: Clearly present the hidden logical relationships, processes, or structures within linear text.

• Clarify Thoughts: Restructure information through visual expression during your conversation with the AI, helping you "think through" complex problems.

• Build Visual Assets: Generated images are automatically tagged with metadata and become searchable, reusable parts of your personal knowledge base.

Getting Started

Prerequisites

• Model Configuration: This feature relies on specific AI models. Ensure your Base URL setup is ready. remio currently uses GPT Image 2 as the default hosted image model; we may change this default as quality and speed improve. If you use BYOK, pick a provider that can generate images.

• Credit Usage: For the first two weeks after launch, you get 10 free generations per day. Under normal use, generating one image consumes about 20 times the credits of a standard text-based conversation.

How to Start

Click the ➕ icon next to the input box and select "Create Image" to enter the image generation mode . The placeholder text in the input box will change to "Describe the image you want to generate" to guide you.

You can also type /image in Chat (include the trailing space) to start the same workflow.

Core Features

Generating an Image

1. After entering "Generate Image" mode, describe the image you want in the input box using natural language.

2. You can also upload a local image, which the AI will use as a reference or a subject for modification.

3. Click send, and the AI will begin generating the image. A loading state will be displayed, and you can click Cancel at any time to stop the process.

Image Card Actions

Once generated, the image appears as a card in the conversation, offering the following actions:

Continuous Conversation and Fine-Tuning

You can request changes to the generated image directly in the conversation, such as "Change the background to a dark color" or "Add a decision node to the flowchart." The AI will understand your request and produce a new version.

Advanced Tips & Best Practices

The AI automatically selects the best visual format based on your description. To get more precise results, you can optimize your prompts using these suggestions:

Best Practices:

• Focus on Structure, Not Aesthetics: Your description should prioritize clear logic and relationships over artistic beauty. The goal of remio's image generation is to aid understanding, not to create commercial illustrations.

• Leverage Reference Images: When you need a specific style or want to modify an existing image, uploading a reference image helps the AI better understand your requirements.

Troubleshooting & FAQ

Q: Why isn't anything happening after I click "Generate"?

A: Check that image generation is available for your current model setup. remio's default chat path includes a hosted image model (currently GPT Image 2). If you use BYOK, switch to a provider that supports image generation.

Q: Can I upload an image for the AI to edit?

A: Yes, this is supported. You can upload an image, and the AI will use it as a subject or reference to be modified or expanded upon, or as a reference for style and structure.

Q: How do I save a generated image to my notes?

A: The current version does not support saving images directly to a note with one click. You can first use the "Save image as..." function to download it locally, then manually add it to your notes. This feature is on our roadmap.

Q: My image generation failed with a "Moderation" error. What does that mean?

A: This means your prompt likely triggered a content safety filter. Please try rephrasing your request using more neutral, professional language focused on knowledge expression.