Adjusting Models & Appearance

Learn how to switch between local and cloud-hosted models, change the aesthetics and layout of your Pieces Copilot chat view, and more.

Managing the LLM Runtime

There are dozens of local and cloud-hosted LLMs to choose from within the Pieces Desktop App—here’s how to do it.

LLM Runtime Modal

At the bottom left of the Pieces Copilot view is the active model. By default, this is a cloud LLM.

Clicking this button opens the Manage Copilot LLM Runtime modal, where you can enter your own API key or select local and cloud-hosted LLMs served through Pieces.

Some cloud models are available to Pieces Pro users only (for example: OpenAI GPT-5.2 Pro/GPT-5.2, Anthropic Claude 4.5 Opus, Google Gemini 3 Pro Preview). To unlock these premium models, see [Pieces Pro](/paid-plans).

Model Selection

Pieces Copilot allows you to choose between cloud-hosted models and on-device models.

Each option has its benefits:

  • Cloud LLMs: Typically offer state-of-the-art performance and are ideal for complex queries requiring deep context. Supported providers and example models include OpenAI (GPT-5.2 Pro, GPT-5.2, GPT-5.1, GPT-5 Thinking, GPT-5, GPT-5 Fast, o4 Mini, o3 Pro, o3 Mini, o3, o1, GPT-4.1, GPT-4o, GPT-4o Mini), Anthropic (Claude 4.5 Opus, Claude 4.5 Sonnet, Claude 4.5 Haiku, Claude 4 Sonnet, Claude 3.7 Sonnet, Claude 3.5 Sonnet, Claude 3.5 Haiku), and Google (Gemini 3 Pro Preview, Gemini 3 Flash Preview, Gemini 2.5 Pro, Gemini 2.5 Flash, Gemini 2.5 Flash Lite, Gemini 2 Flash Lite). See the complete list on the Cloud Models page.

  • On-Device LLMs: Ensure data privacy and are optimal for offline or air-gapped environments.

You can read this documentation containing all local and cloud-hosted LLMs available through the Pieces Desktop App.

Resetting Conversations

In case you need a fresh start or want to clear the current context, the interface includes options (accessible via the Chat Options menu) to reset the conversation.

This is particularly useful when you want to switch focus or change the conversation pipeline.

Search Functionality

To browse and download local models, click the Active Model button and hover over All Models, then scroll down to find available local models.

Local models are downloaded on-demand through PiecesOS. Simply select a local model to download and use it.

Local models run entirely on your device through PiecesOS, providing complete privacy and offline functionality without any external dependencies.

Adjusting Chat Appearance

Within the same modal area, a Settings Gear icon gives you access to personalization options.

From here, you can choose a chat accent color to customize the look and feel of your Copilot interface, and enable or disable the option to enable LTM by default when starting a new chat.

When enabling Long-Term Memory Context by default, every new chat automatically incorporates your saved long-term memory context, ensuring that your conversations are always informed by your previous work.

You can also use the keyboard shortcut cmd+shift+t (macOS) or ctrl+shift+t (Windows/Linux) to toggle the Dark/Bright theme for the Pieces Desktop App.