Create design files with Open Design

Olares version: 1.12.7+▾

Olares 1.12.7+Olares 1.12.6

Open Design is an open-source AI design studio that turns natural-language briefs into working design files. Its built-in OpenCode agent uses a model provider you configure to create previewable HTML prototypes, landing pages, dashboards, wireframes, and slide decks.

On Olares, you can connect Open Design to a local model through Router or bring an API key for a cloud provider. Projects keep their prompts, references, previews, and generated files together so you can iterate on a design before exporting it.

Current output limitation

Open Design 0.22.1 on Olares does not generate standalone images from text. Use it to create pages, prototypes, dashboards, or slide decks instead.

Learning objectives

In this guide, you will learn how to:

• Install Open Design on Olares.
• Connect a local model through its OpenAI-compatible API.
• Create, refine, and export a design project.
• Troubleshoot common model connection and generation issues.

Prerequisites

Before you begin, you need:

• An Olares device running Olares 1.12.7 or later, with sufficient disk space and memory.

• Access to Router for local model connections.

• The following model:

  Model type	Model	How to get it
  Chat	Qwen3.8-27B (llama.cpp)	Install from Market

Optional: Use a different model

Install another chat model from Market, or create a model instance with Engine Base apps. Then select it as the default chat model on Router's Default models page. Clients configured with default-chat use the new default without changing their model settings.

To keep a client on one specific model, use the full model name from Router's How to call this model window instead.

Install Open Design

1. Open Market and search for "Open Design".
2. Click Get, then Install, and wait for installation to complete.

Connect a local model

Get the model connection details

How model connections work

On Olares 1.12.7 and later, AI clients connect through Olares Router. Router provides the Base URL and routes default-chat to the model selected on its Default models page.

This guide uses Qwen3.8-27B (llama.cpp) as the default chat model. For connection formats and API key requirements, see Connect AI apps.

In this guide, Open Design connects to Qwen3.8-27B (llama.cpp) using the OpenAI-Compatible API format:

1. Open Router from Launchpad. On Default models, set Qwen3.8-27B (llama.cpp) as the default chat model.

2. Go to LLM, find Qwen3.8-27B (llama.cpp), and click View connection example on its row.

3. In How to call this model, select Apps in Olares and copy the Base URL, including /v1.

4. For the model name, use default-chat.

   Selecting a model in your app

   default-chat is a routing name and does not appear in model lists loaded from Router. If the app you are connecting to Router loads a model list, add default-chat manually. If the app only supports selecting a listed model, select the full model name shown in Router's How to call this model window.

5. For the API key, leave the field empty. If the app you are connecting to Router requires a value, enter olares.

Add the model to Open Design

1. Open Open Design from Launchpad.

2. Open Settings. On the Models & providers page, select API provider.

3. Under Provider preset, select Custom provider.

4. Configure the provider:

   • Base URL: Paste the Base URL copied from Router, including the trailing /v1.
   • API key: Enter olares.
   • Model: Enter default-chat.
   • Max tokens (optional): Enter 65536. You can use 32768 for a smaller context window.

5. Click Test to verify the connection. A green Connected status confirms that the connection is established. Changes are saved automatically.

Start with a small test

Ask Open Design to create a single title slide before starting a multi-page deck or landing page. A successful preview confirms that the Base URL, model name, and token limit work together.

Create a design project

1. On the Home page, select a project type, such as Prototype, Slide deck, Document, or Website clone.

2. Choose a Skill preset that matches your task, such as Blog Post. You can also open one of the examples below the prompt to use it as a starting point.

3. Optionally, select a Design system and Working directory.

4. Enter a specific brief. Include the intended audience, content, visual direction, and output format. For example:

   Create a title slide for a product launch presentation.
   Use a dark background, a bold geometric title, and a teal accent.
   Include the product name "Northstar" and the subtitle "Find your next move."

5. Check that the intended model is selected, then click the arrow to start generating.

6. Wait for the workspace to open and the preview to render.

7. Refine the result in the conversation. You can drag in a reference image or type @ to attach a project file, then describe the part you want to change.

8. When the result is ready, open the export menu and download it as ZIP, PDF, or PPTX, depending on the project type.

Optional: Connect a cloud model

You can use a cloud provider instead of the local model. Open Settings, select API provider on the Models & providers page, then choose a provider preset or configure a custom provider with the matching protocol, Base URL, API key, and model.

Provider	Protocol	Base URL
OpenAI	OpenAI	https://api.openai.com/v1
OpenRouter	OpenAI	https://openrouter.ai/api/v1
DeepSeek	OpenAI	https://api.deepseek.com
Anthropic	Anthropic	https://api.anthropic.com
Qwen	OpenAI	https://dashscope.aliyuncs.com/compatible-mode/v1
SiliconFlow	OpenAI	https://api.siliconflow.cn/v1
Kimi	OpenAI	https://api.moonshot.cn/v1

Cloud usage is billed by the selected provider. Use that provider's exact model name and a valid API key.

Troubleshooting

The connection test fails or returns 404

Copy the Base URL from Apps in Olares in Router's How to call this model window, including /v1. Check that the default chat model shows Callable on LLM.

Generation stops before the design is complete

Set Max tokens to 65536 or 32768. Multi-page slide decks need more output tokens than a single page or cover slide.

Open Design does not respond

Return to the model app's Model Console. Continue only when Model shows READY and Engine shows RUNNING.

Open Design uses the wrong model

Use default-chat and check its target on Router's Default models page. To pin a specific model, copy its full name from Router instead.

Open Design does not create a standalone image

This is a limitation of Open Design 0.22.1 on Olares. Select a Skill for a page, slide deck, dashboard, wireframe, or interactive prototype instead.

Learn more

• Host local large language models with Engine Base apps: Deploy local models with Engine Base apps and connect them through Router.
• Connect AI apps to model services: Understand model names, API formats, and Base URLs.
• Open Design website: Explore the upstream project and its capabilities.