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

# Chat

> Interact with AI for biology research through chat

Conduct research through natural conversation with OpenBio's AI chat interface. Ask questions, run analyses, query databases, and get insights—all through an intuitive chat that understands scientific context and maintains conversation history.

<Info>
  **Starting a New Chat**
  Click the **"+"** button in the top right corner of the chat panel to begin a new research conversation within your current project.
</Info>

## Sending Messages

### Basic Message Types

<Tabs>
  <Tab title="Questions">
    Ask anything about biology, research methods, or scientific concepts:

    ```
    What is the role of p53 in cancer development?
    How does CRISPR-Cas9 target specific DNA sequences?
    What are the latest FDA-approved kinase inhibitors?
    ```
  </Tab>

  <Tab title="Database Queries">
    Request searches in biological databases:

    ```
    Search UniProt for human membrane proteins
    Find PDB structures of HIV protease
    Query ChEMBL for EGFR inhibitors with good bioavailability
    ```
  </Tab>

  <Tab title="Analysis Requests">
    Ask for analyses and predictions:

    ```
    Predict the structure of this protein sequence: MKTAYIAK...
    Analyze the binding affinity of this ligand to the target protein
    ```
  </Tab>

  <Tab title="File References">
    Include uploaded files in your requests:

    ```
    Analyze the sequences in @protein_data.fasta
    Show me the structures of @model1.pdb and @model2.pdb
    Summarize the findings in @research_paper.pdf
    ```
  </Tab>

  <Tab title="Workflow Execution">
    Run saved workflows for automated analyses:

    ```
    Run /protein-analysis on this sequence
    Execute the /literature review workflow for "CRISPR gene editing"
    ```
  </Tab>
</Tabs>

### Message Controls

**Sending Messages**:

* Type your message in the input box at the bottom
* Press **Enter** to send (single line)
* Press **Shift + Enter** for new lines without sending
* Click the **arrow button** to send

**During OpenBio Response**:

* **Stop generation**: Click the red stop button
* **Scroll up**: Read previous messages without interrupting

## Tool Calls and Visualizations

<Info>
  **What are Tool Calls?**
  When OpenBio needs to perform actions like searching databases or running predictions, it uses specialized tools. Each tool call appears as an expandable card showing what was done.
</Info>

**Tool call components**:

* **Tool name**: Which service was used (e.g., "PDB Search", "UniProt Query")
* **Parameters**: What was searched for or analyzed
* **Status**: Running, completed, or failed
* **Actions**: Options to view details or save results, and view the visualization

## Managing Conversations

### Chat History

<Info>
  **Location**: Chat history button in the top right corner of the chat panel
</Info>

List of all your chats in your current project.

### Search and Organization

**Search Chats**:

* Click the search box at the top of the chat history list
* Search across chat with keywords

**Pin Important Chats**:

* Hover over any chat and click the pin icon
* Pinned chats stay at the top regardless of activity
* Perfect for ongoing research projects

**Rename Chats**:

* Click the chat title in the header to edit
* Or right-click chat in chat history and select "Rename"
* Use descriptive names for easy identification

**Delete Chats**:

* Hover over any chat and click the delete icon
* Deleted chats cannot be recovered

<Warning>
  Deleted chats cannot be recovered.
</Warning>

## Advanced Features

### File References

Include your research files directly in conversations:

**Method 1: @ Mention**

1. Type `@` in your message
2. Select from autocomplete suggestions
3. File appears as a blue chip in your message

**Method 2: Using attach button in the chat input**

1. Click the attach button in the chat input
2. Select from autocomplete suggestions
3. File appears as a blue chip in your message

**Method 3: From the files menu**

1. Right-click any file in the files panel
2. Select "Add to Chat"
3. File reference inserted

**How it works**: AI reads and analyzes the file content, incorporating it into the response.

### Workflow Integration

Execute saved research workflows:

**Method 1: Type / in the chat input**

1. Type / in the chat input
2. Select from autocomplete suggestions
3. Workflow appears as a purple chip in your message

**Method 2: Using workflow button in the chat input**

1. Click the workflow button (/) in the chat input
2. Select from autocomplete suggestions
3. Workflow appears as a purple chip in your message

Read more about file references in the [File Management](https://openbio.ai/user-guide/features/file-management) section and [Workflows](https://openbio.ai/user-guide/features/workflows) section.

### Tool Suggestions

Suggest specific tools to use in your conversation:

**Method 1: Type ! in the chat input**

1. Type `!` in your message
2. A dropdown appears with available tools
3. Type to search by tool name, display name, or description
4. Select a tool from the suggestions
5. Tool appears as a chip in your message

**Method 2: Using tool button in the chat input**

1. Click the tool button (!) in the chat input toolbar
2. A dropdown appears with all available tools
3. Search or browse tools by category
4. Select a tool from the list
5. Tool appears as a chip in your message

**How it works**: When you suggest a tool, OpenBio will prioritize using that tool for your request. This is helpful when you want to use a specific tool or ensure OpenBio uses a particular capability.

**Example**:

```
!search_uniprot_tool Find human insulin sequences
```

This tells OpenBio to use the UniProt search tool specifically for finding insulin sequences.

<Info>
  Tool suggestions help guide OpenBio to use specific tools, but OpenBio may still use other tools if needed to complete your request effectively.
</Info>

### Model Selection

Choose the right AI model for your task:

<Info>
  **Location**: Model selector in bottom-left of chat input
</Info>

**Quick Selection Modes**:

* **Fast** - Quick responses for simple queries and database searches
* **Balanced** - Optimal for most research tasks (recommended)
* **Max** - Highest quality for complex analysis, predictions, and design tasks

**Available Models**:

* **Claude Sonnet 4**: Best overall performance, complex reasoning (use for predictions and complex reasoning tasks)
* **Gemini Flash 2.5**: Fast responses, general research tasks
* **Other models**: Additional options for specific needs

**Changing Models**:

1. Click current model name or mode (Fast/Balanced/Max)
2. Select from dropdown menu
3. Next message uses the new model
4. Selected model persists for all future messages

**Adding Custom Models**:
Access Settings → Add Model to configure additional AI models for specialized workflows.

### Code Mode (Data Analysis)

Activate Code Mode for Python-based data analysis and computational tasks:

<Info>
  **Location**: Code icon (\</>) button in chat input toolbar
</Info>

**What is Code Mode?**
Code Mode tells OpenBio to prioritize using Python code execution tools for computational analysis, data visualization, and scientific computing.

Watch a demo of Code Mode:

<iframe className="w-full aspect-video rounded-xl" src="https://www.youtube.com/embed/sds1Eiz0BhA" title="Code Mode Demo" frameBorder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" referrerPolicy="strict-origin-when-cross-origin" allowFullScreen />

**Activating Code Mode**:

1. Click the code icon (\</>) in the chat input area
2. Icon and the chat input area highlights in purple when active
3. Your preference is saved automatically

**When to Use Code Mode**:

* Data analysis and visualization
* Statistical calculations
* Scientific computing tasks
* Bioinformatics sequence analysis
* Creating plots and charts

**Features**:

* Stateful Python environment (variables persist across executions)
* Pre-installed data science libraries (pandas, numpy, matplotlib, etc.)
* File upload/download integration with your project
* Rich output support (plots, DataFrames, images)

Read more in the [Data Analysis (Code Mode)](/user-guide/features/data-analysis) guide.

### Best Practices for Chat

<AccordionGroup>
  <Accordion title="Be Specific">
    Instead of "tell me about proteins," ask "What are the key differences between fibrous and globular proteins in terms of structure and function?"
  </Accordion>

  <Accordion title="Provide Context">
    AI remembers conversation history. Reference previous results: "Now compare that structure to human insulin and identify the differences."
  </Accordion>

  <Accordion title="Break Down Complex Tasks">
    For multi-step analyses, complete one step at a time. Verify results before proceeding to the next step.
  </Accordion>

  <Accordion title="Be descriptive">
    Be descriptive in your requests. Write all the information that you think is relevant to the task.
  </Accordion>

  <Accordion title="Include Your Data">
    Upload your files and reference them with @. AI can analyze sequences, structures, and other data directly.
  </Accordion>

  <Accordion title="Ask Follow-up Questions">
    Don't hesitate to ask for clarification, more details, or alternative approaches.
  </Accordion>
</AccordionGroup>

## Troubleshooting

<AccordionGroup>
  <Accordion title="Messages won't send">
    **Solutions**:

    * Check internet connection
    * Ensure message isn't empty
    * Try refreshing the page
  </Accordion>

  <Accordion title="AI not responding">
    **Solutions**:

    * Wait 30 seconds (complex queries take time)
    * Check if stop button is available
    * Try regenerating the response
    * Start a new chat if issues persist
  </Accordion>

  <Accordion title="Can't find old messages">
    **Solutions**:

    * Scroll up in the chat area
    * Use search to find specific content
    * Check if you're in the right chat
  </Accordion>

  <Accordion title="File references not working">
    **Solutions**:

    * Ensure file is uploaded to current project
    * Type @ to see autocomplete suggestions
    * Check file name spelling
    * Refresh page if autocomplete isn't working
  </Accordion>

  <Accordion title="Tool suggestions not appearing">
    **Solutions**:

    * Type `!` to trigger tool suggestions dropdown
    * Click the tool button (!) in the chat input toolbar
    * Check internet connection (tools are loaded from server)
    * Refresh page if dropdown doesn't appear
    * Try typing part of the tool name to search
  </Accordion>
</AccordionGroup>

## Next Steps

<CardGroup cols={2}>
  <Card title="File Management" href="/user-guide/features/file-management">
    Upload and organize research files
  </Card>

  <Card title="Workflows" href="/user-guide/features/workflows">
    Create reusable chat workflows
  </Card>

  <Card title="Predictions" href="/user-guide/features/predictions">
    Submit AI prediction jobs
  </Card>

  <Card title="Notebooks" href="/user-guide/features/notebooks">
    Take notes from conversations
  </Card>

  <Card title="Data Analysis (Code Mode)" href="/user-guide/features/data-analysis">
    Use Python for data analysis and visualization
  </Card>

  <Card title="MCP Support" href="/user-guide/features/mcp-support">
    Connect external tool servers
  </Card>

  <Card title="Zotero Support" href="/user-guide/features/zotero-support">
    Integrate your reference library
  </Card>
</CardGroup>

<Tip>
  **Pro Tip**: Start each chat with a clear, specific first message. This sets the context and helps OpenBio provide more relevant responses throughout your conversation.
</Tip>