Skip to main content
Docs
Your First Chat

Your First Chat

Learn how to use the Elyxir Studio chat interface and interact with AI models.

The Elyxir Studio chat interface provides a powerful way to interact with AI models directly in your browser. This guide walks you through your first conversation.

Accessing Studio Chat

  1. Sign in to studio.elyxir.ai
  2. Select your organization from the dropdown
  3. Click Studio in the main navigation

The Chat Interface

The Studio chat interface consists of:

  • Message Input: Type your prompts at the bottom of the screen
  • Conversation History: View past messages in the main area
  • Model Indicator: See which model is responding to your request
  • Settings Panel: Configure model preferences and options

Starting a Conversation

  1. Click in the message input field
  2. Type your message or question
  3. Press Enter or click the send button

Understanding Alvin Routing

Alvin analyzes your prompt and routes it to the optimal AI model based on:

  • Task Type: Code generation, creative writing, analysis, etc.
  • Complexity: Simple queries vs. complex reasoning tasks
  • Context Length: Short vs. long conversations
  • Cost Efficiency: Balancing quality with cost

When Alvin routes your request, you'll see which model was selected in the response header.

Selecting a Specific Model

To use a specific model instead of Alvin's automatic selection:

  1. Click the Model dropdown in the chat settings
  2. Browse or search for your preferred model
  3. Select the model (e.g., "GPT-4o", "Claude 3.5 Sonnet")
  4. Your next message will use that model

Chat Features

Multi-turn Conversations

Continue the conversation naturally - the AI remembers previous messages:

You: What is TypeScript?
AI: TypeScript is a typed superset of JavaScript...

You: How do I install it?
AI: You can install TypeScript using npm...

Web Search Integration

Some models support real-time web search:

  1. Enable Web Search in the chat settings
  2. Ask questions that require current information
  3. The AI will search the web and cite sources

Code Highlighting

Code in responses is automatically syntax-highlighted:

function greet(name) {
  return `Hello, ${name}!`;
}

Chat Settings

Access settings via the gear icon or settings panel:

SettingDescription
ModelSelect a specific model or use Alvin routing
TemperatureControl response creativity (0-1)
Max TokensLimit response length
Web SearchEnable real-time web search

Keyboard Shortcuts

ShortcutAction
EnterSend message
Shift + EnterNew line in message
Ctrl/Cmd + KOpen model selector
EscClose dialogs

Managing Conversations

New Conversation

Click New Chat or use the keyboard shortcut to start fresh.

Conversation History

Access previous conversations from the sidebar. Conversations are saved automatically.

Exporting Chats

Export your conversation for reference:

  1. Click the menu icon on a conversation
  2. Select Export
  3. Choose your format (Markdown, JSON)

Tips for Better Results

  1. Be specific: Clear, detailed prompts get better responses
  2. Provide context: Include relevant background information
  3. Iterate: Refine your prompt based on responses
  4. Use system prompts: Set the AI's behavior for the conversation

Example Prompts

Code Help

Write a Python function that validates email addresses using regex. Include error handling and docstrings.

Analysis

Analyze the pros and cons of microservices vs monolithic architecture for a startup with 5 developers.

Creative Writing

Write a professional email declining a meeting request while suggesting an alternative time.

Next Steps

Your First Chat | elyxir