> ## Documentation Index
> Fetch the complete documentation index at: https://docs.getspine.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Chat Block

> Conversational interface for interacting with any AI model

## Features

* **300+ Models Available**: Switch between GPT-5, Claude, Gemini, Grok, Llama, and hundreds more
* **Conversation Branching**:  Click the branch button under any message to start a new conversation thread from that point. Your previous thread will be preserved
* **Context Sharing**: Automatically shares conversation context with connected blocks (both in and out)
* **Model Switching**: Change models mid-conversation to get different perspectives
* **Web and Doc Search**: Specify if you want the chat block to search through the web and/or all "documents" anywhere on the canvas (includes files, YouTube videos, webpages). Note: This is for broader web and document search — if you connect specific blocks to the chat block, they will be used as context regardless of these settings
* **Multi-turn Conversation** — Supports iterative exploration, brainstorming, and back-and-forth refinement within a single block

## How to Use

1. **Create a Chat Block**: Double click on canvas → Select "Chat Block"
2. **Change Your Model**: Click the model dropdown to choose your preferred AI model
3. **Start Conversing**: Type your message and press Enter
4. **Branch Conversations**: After sending a message in the chat, you can click the "Branch" button to explore different directions while preserving all chat history and context
5. **Connect to Other Blocks**: Drag connection lines to share context with other blocks

## Best Practices

* **Switch models for different strengths**: Use OpenAI models for creative tasks, Claude for coding and analysis, Gemini 2.5 for coaching, Grok 4 for brainstorming
* **Branch conversations** when you want to explore multiple directions without losing your main thread
* **Connect to other blocks** to transform chat insights into research, visuals, or structured outputs

## Video Demo