Generating Images — Create Visuals Directly in Chat

Askimo lets you generate images without leaving the chat window. Depending on the AI provider and model you have configured, images are produced either natively inline (the model responds with an image directly) or through a dedicated image model you select yourself (such as DALL-E 3).

---

How Image Generation Works

Askimo supports two image generation paths:

Native In-Chat Generation

Some models — such as Gemini Banana — can produce images as part of a normal chat response. When Askimo detects this capability automatically at startup, it activates native mode: you simply describe what you want in plain text and the model returns an image alongside any accompanying explanation.

In native mode the Image toggle button is hidden from the toolbar because switching modes is unnecessary — the model handles everything on its own.

Dedicated Image Model

Most providers separate their image generation capability into a specialist model (for example, OpenAI’s DALL-E 3 or Anthropic’s Claude Sonnet). In this case you click the Image icon (📷) in the chat input toolbar to enter Image creation mode, type your prompt, and send it. Askimo routes the request to the configured image model and displays the result inline.

Nota

The Image icon only appears when your active provider has a dedicated image model configured and the current chat model does not support native in-chat generation.

---

Configuring Your Image Model

Before you can use the dedicated image model path, make sure the right model is set for your provider.

To set the image model:

1. Open Settings with Ctrl + , (or via the menu bar)
2. Navigate to the AI Providers tab
3. Select the provider you use for image generation
4. Locate the Image Model field and click it to open the model picker
5. Choose from the list or type a model name, then confirm

Cuidado

There is no default image model — you must explicitly set one for each provider before image generation becomes available.

---

Generating an Image

With a Dedicated Image Model

1. Start or open a chat session with a provider that has an image model configured
2. Click the Image icon (📷) in the chat toolbar — a Image mode chip appears near the input field to confirm you are in image creation mode
3. Type your image prompt (be descriptive — include style, colours, composition, and mood for best results)
4. Press Enter or click Send
5. Askimo sends your prompt to the image model and displays the result directly in the conversation

To return to normal chat, click the ✕ on the Image mode chip or click the Image icon again.

With a Native Multi-Modal Model

1. Switch to a model that supports native image generation (e.g., gemini-banana on Google Gemini)
2. Type your image request as a regular message — for example: “Draw a minimalist logo of a mountain at sunset”
3. Send the message; the model replies with the generated image inline

Dica

Native multi-modal models can combine images with text in a single response. Ask for an image and an explanation in the same prompt to get both at once.

---

Writing Effective Image Prompts

Good prompts produce better images regardless of the underlying model. Keep these guidelines in mind:

• Be specific about subject and setting — “a tabby cat sitting on a windowsill in soft morning light” beats “a cat”
• Mention an art style — “watercolour illustration”, “photorealistic 4K”, “flat vector art”
• Specify aspect ratio or composition when relevant — “wide landscape”, “portrait format”, “close-up”
• Include mood or atmosphere — “calm, pastel palette”, “dramatic lighting”, “cyberpunk neon glow”

---

Troubleshooting

The Image icon is not visible in the chat toolbar → Either the active model supports native in-chat generation (so no toggle is needed), or no image model has been configured for the current provider. Go to Settings → AI Providers and set an image model.

I get an error about a missing image model → The Image icon is shown but no image model is saved for your provider. Open Settings → AI Providers, select your provider, and fill in the Image Model field.

The generated image looks wrong or the quality is poor → Refine your prompt using the tips above. For DALL-E 3, longer and more detailed prompts generally yield better results. For local models, check that the model is fully downloaded and that your hardware meets the requirements.

Images are not appearing inline in the chat → This can happen with very large base64 responses on slower machines. Wait a moment for rendering to complete; if the issue persists, try switching to a smaller or cloud-hosted image model.