AI Chat

Indigo’s AI chat lets you ask questions across your entire meeting history using the AI model of your choice. Instead of scrolling through transcripts, you query in natural language and get context-aware answers with citations.

Model-agnostic architecture

Indigo does not lock you into a single AI provider. You choose which model powers your chat experience.

Supported providers include:

• OpenAI (GPT-4, GPT-4o, etc.)
• Anthropic (Claude)
• Google (Gemini)
• xAI (Grok)

You can switch between providers at any time. Your meeting data stays the same — only the AI model processing your queries changes.

Bring Your Own Key (BYOK)

Indigo uses a BYOK model for AI chat. You provide your own API key from your preferred provider, and Indigo routes queries through that provider’s API.

This means:

• You control costs — usage is billed directly by your AI provider
• You control data flow — queries go to the provider you trust
• No vendor lock-in — switch providers without losing any meeting data

Tip

Configure your API keys in Indigo’s settings. You can also manage keys through the Indigo CLI using indigo config.

Context-aware responses

AI chat operates across Indigo’s hot and cold data architecture:

Data layer	What it includes
Hot data	Active meetings, live transcripts, current decision streams
Cold data	Historical transcripts, past decisions, action item archive, institutional knowledge

When you ask a question, the AI searches across both layers and returns answers grounded in your actual meeting content. Responses include citations — links back to the specific transcript segments, decisions, or action items that informed the answer.

Querying across meetings

AI chat is not limited to a single meeting. You can query across your full history:

• “What did we decide about the pricing model last quarter?”
• “Who committed to the API integration and when is it due?”
• “Summarize all customer feedback from the last 30 days of calls”
• “What blockers has the engineering team raised in the last 3 standups?”

The AI pulls from all relevant meetings, cross-references decisions and actions, and synthesizes a response with citations to the source material.

How it differs from generic AI

Generic AI chat	Indigo AI chat
No context about your meetings	Grounded in your transcript data
Requires copy-pasting transcript text	Searches across all meetings automatically
Single-session memory	Persistent cross-meeting memory
General-purpose responses	Meeting-specific, citation-backed answers

Note

AI chat does not require prompt engineering. Ask questions in plain language — Indigo handles the context retrieval and prompt construction behind the scenes.