An AI-native workspace built around Supertags—a system inspired by object-oriented programming that turns unstructured notes into structured, queryable records. Unlike folder-based tools, Tana’s knowledge graph mirrors associative thinking, and Supertags let you define types (Meeting, Task, Decision, etc.) with fields and AI instructions, so every note becomes data an agent can filter and act on. Its meeting agent records and transcribes meetings (5000 credits/month on Pro ≈ 22 × 30-min meetings), captures outcomes as proposals for your review, and syncs with Google Calendar. A hosted MCP server (home.tana.inc/mcp) lets Claude Code, Claude Desktop, and any MCP client read and write your Tana content—writes go through the same proposal review as the built-in AI. Pricing: Free (500 credits/month, no meeting agent), Plus $8/mo** (2000 credits, meeting notes, voice transcripts), **Pro $14/mo (5000 credits, full AI workflows). Note: tana.app is a separate, unrelated inventory management product—do not confuse the two.
Pros & Cons:
Pros: Supertags are a genuinely different approach to knowledge management—they make notes structured and queryable without forcing you into rigid database schemas, and AI instructions on each type keep entries consistent; the MCP server lets external AI agents (Claude Code, Cursor) read and write your knowledge graph, with writes gated by proposal review, which is safer than direct writes; the meeting agent captures outcomes as structured proposals rather than raw transcripts, and pinning documents to recurring meetings updates them in place instead of creating duplicates; voice-first workflows (mobile voice chat, meeting recording) are well-integrated, and the 2026 updates added natural-language search, daily Slack briefs, and a multitasking voice agent that can act in the background during meetings.
Cons: The product line has split—tana.inc (Outliner) and tana.app (inventory) are completely different tools, and the brand confusion can mislead users; Free tier does not include the meeting agent, so the core AI value requires Plus ($8/mo) or Pro ($14/mo); AI credits are metered and do not roll over—Pro’s 5000 credits cover about 22 half-hour meetings or 25 hours of live transcription, and heavy meeting use can exhaust them; the learning curve is real—Supertags and the graph model require understanding a different mental model than Notion or Obsidian, and the product surface is broad (meetings, voice, MCP, agents, search); independent third-party validation is thin—most evidence comes from Tana’s own documentation and blog, with limited G2/Capterra reviews surfaced.
Who is it for?:Professionals and teams who want a knowledge base that is structured by design (Supertags) and can be read/written by external AI agents via MCP; heavy meeting-goers who want AI to capture outcomes as structured proposals and sync with calendars; users willing to learn a new mental model in exchange for more powerful querying and automation than Obsidian or Mem; teams exploring “agent memory” that is permissioned and current rather than a stale vector store. Not suited for users seeking a simple, free note-taking app (Obsidian is better), those who want a flat folder structure, or anyone who might confuse it with the tana.app inventory tool.