Tana

Tana — Technical Overview

Tana is an agentic meeting platform that embeds AI agents directly into native video calls, executing real work in real time — without a bot joining the call. It is LLM-agnostic, routing across Claude, Gemini, and OpenAI models, and exposes a public API alongside MCP (Model Context Protocol) support for extensible, programmatic integration.

Key architectural characteristics:

• Botless capture: Meeting intelligence is gathered natively — no third-party bot participant is injected into the call session.
• Real-time agent execution: Agents can file issues to Linear/Jira/GitHub, update HubSpot CRM records, and draft/send Slack messages or emails before the call ends — not in a post-meeting batch.
• Shared knowledge graph: All meeting data is indexed into a team-wide knowledge graph, queryable across sessions.
• LLM-agnostic runtime: Swap or combine Claude, Gemini, and OpenAI models per workflow.
• MCP + public API: Integrate Tana into existing toolchains or build custom agents on top of meeting context.
• Compliance posture: SOC 2 Type 2 and GDPR compliant; user data is never used for model training.

Integrations: Linear, Jira, GitHub, HubSpot, Slack, Google Calendar, Outlook, Claude Code, Codex, Cursor, Gemini, Lovable, MCP. Google Workspace and Microsoft 365 integrations are forthcoming.

Pricing: Freemium. Free tier (5 hosted meetings/month, 50 AI queries). Pro at $20/user/month early bird (yearly). Max at $80/user/month early bird (yearly). Business at custom pricing.