# Getting Started with AI

Weavestream includes a built-in AI chat that can analyze your data, answer questions, identify patterns, and generate summaries. You choose which AI provider to use based on your needs — cloud-based for maximum capability or local for privacy.

# Choosing an AI Provider

Weavestream supports three AI providers. You only need to set up one to get started.

# Claude (Anthropic)

The default and most capable option. Claude runs in the cloud via Anthropic's API.

Pros:

• Most accurate and capable analysis
• Handles large, complex datasets well
• Excellent at following nuanced instructions

Setup:

1. Open Settings (Cmd+Comma) and go to Intelligence
2. Click on Claude in Weavestream
3. Enter your Anthropic API key (starts with sk-ant-)
4. Click Save API Key

Don't have an API key? There's a link in the settings to get one from Anthropic's website.

# Ollama (Local)

A privacy-first option that runs AI models entirely on your Mac. No data leaves your device.

Pros:

• Complete privacy — data stays local
• No API costs
• Works offline

Considerations:

• Requires Ollama to be installed and running on your Mac
• Quality depends on the model you choose
• May be slower than cloud options, especially for large datasets

Setup:

1. Install Ollama from ollama.ai
2. Download a model (e.g., ollama pull llama3)
3. In Weavestream, open Settings → Intelligence
4. Click on Ollama in Weavestream
5. Enter the server URL (default: http://localhost:11434)
6. Click Test to verify the connection
7. Select a model from the dropdown (click Refresh to discover available models)
8. Adjust Max Context Tokens if needed (default is fine for most uses)
9. Click Save Configuration

# Apple Intelligence

Apple's built-in intelligence is used for specific features like smart chat naming (automatically titling your conversations). It works alongside your primary provider rather than replacing it.

# Setting Your Active Provider

After configuring one or more providers, set your active one:

1. Go to Settings → Intelligence
2. Click Set as Active next to the provider you want to use

The active provider is shown with a green "Active" badge.

# Opening the AI Chat

The AI chat panel appears below the detail view in the main window. You can toggle the chat bubble display in Settings → General.

Select an item, endpoint, or Smart Filter in the sidebar before starting a chat — this tells the AI what data to analyze.

# Your First Analysis

1. Select a source, endpoint, or Smart Filter in the sidebar
2. In the AI chat panel, type a question like "Summarize the key trends in this data"
3. Press Return or click the Send button

Weavestream will send relevant data to the AI along with your question and display the response.

You can also use the quick-start buttons that appear in a new chat:

• Summarize — Get a high-level overview
• Critical Issues — Identify the most urgent items
• Trends — Spot patterns over time
• Action Items — Get a list of recommended next steps

# Next Steps

• Chatting with Your Data — Learn how to ask better questions
• Custom Analysis Prompts — Create reusable analysis instructions
• Privacy & AI Providers — Understand what data is sent where