The verdict: can you vibe code Tana?
You can build a web-based note and AI chat subset, but building the cross-platform native desktop call-capture engine and graph database requires months of complex systems engineering.
Tana combines an outliner graph database, real-time multiplayer editing, custom AI agents, and deep desktop OS audio hooking. While an AI agent can rapidly scaffold a Next.js web frontend and basic LLM chat loops, capturing audio streams from active third-party video calls (Zoom, Teams, Meet) without calendar bots via a native desktop shell, combined with a performant bidirectional backlink graph engine, is a massive undertaking for a solo developer.
Estimated effort: 4-6 months of serious part-time engineering
What you can't replicate
- Cross-platform OS audio driver hooks for third-party meeting call capture
- Compound multi-year knowledge graph database optimizations
- Native integrations and OAuth maintenance overhead across a dozen enterprise tools
Founded
2021
Raised
$28.5M
Team
~50
Cheapest paid tier
$0
What Tana does
An agentic meeting platform and AI-native workspace where custom AI agents interact with live calls, extract decisions, and execute tasks across an interconnected knowledge graph.
Core features
- Botless cross-platform desktop meeting audio capture across Zoom, Teams, and Google Meet
- Real-time speech-to-text transcription with speaker diarization
- Relational knowledge graph engine with bidirectional backlinks and supertags
- Custom AI agents with configurable capabilities, schedules, and voice interfaces
- Proposal engine for staging, diffing, and applying AI-generated edits safely
- Multiplayer real-time collaborative document editing workspace
- Model Context Protocol (MCP) and third-party tool integrations (GitHub, Linear, Slack)
The business
Pricing
- Free$0
- Pro (Early Bird)$20/mo
- Max (Early Bird)$80/mo
- BusinessCustom
Funding
$28.5M from Lightspeed Venture Partners, Northzone, Tola Capital, SignalFire
Pay vs build, cumulative
Break-even at month 1 — after that, every month is money kept.
The hard parts of vibe coding Tana
- Capturing system audio and injecting meeting bots across diverse OS-level audio permissions on macOS, Windows, and Linux
- Architecting a performant relational graph engine with typed entity schemas and deep backlinks
- Building a safe proposal preview and diffing system for autonomous AI-driven document mutations
- Orchestrating low-latency real-time voice agent pipelines with multi-participant state handling
How to vibecode Tana
Prerequisites
Node.jsfree
Required runtime for building the full-stack Next.js application.
GitHubfree
Source code control and CI/CD repository hosting.
Anthropic API KeyPay-as-you-go (~$15-$30/mo)
Powers Claude 3.5 Sonnet for agent reasoning and document edits.
AI coding tools
Recommended stack
| Frontend | Next.js with Tailwind CSS and TipTap rich-text editor |
|---|---|
| Backend | Next.js Server Actions and Route Handlers |
| Database | Neon (Serverless Postgres with pgvector for graph relationships and semantic search) |
| Auth | better-auth |
| Payments | Stripe |
| Other | Vercel AI SDK for LLM tool calling and streaming, Deepgram API for real-time audio transcription, Liveblocks for real-time multiplayer document editing, ElevenLabs API for voice agent playback |
Build guide
01Project Scaffolding and Database Schema
Initialize the Next.js application with TypeScript, Tailwind CSS, and Neon Serverless Postgres via Drizzle ORM.
Scramble a new Next.js 16 project with Tailwind CSS, TypeScript, and Drizzle ORM configured for Neon Serverless Postgres. Define the database schema for users, spaces, documents, supertags/fields, meetings, and AI proposal logs. Establish UUID primary keys, foreign key constraints with cascade deletes, and timestamp audit fields across all tables. Set up better-auth for secure credential and session management. Ensure the project structure cleanly separates server actions, database models, and React UI components. Provide clean migration scripts and instructions to verify database connectivity.02Graph Knowledge Engine & Rich-Text Editor
Implement the core document editor with bidirectional backlink tracking and mention pills.
Build a rich-text document editor component using TipTap within Next.js that supports slash commands (headings, lists, code blocks, Mermaid diagrams), inline image attachments, and @mentions that dynamically query and reference other documents. Implement a backend service that automatically parses document content for markdown links and mentions to update a relational backlinks table. Build a sidebar side panel component that queries and renders incoming backlinks grouped by source document and field reference in real-time.03Meeting Recorder and STT Pipeline
Create the meeting capture interface and audio transcription processing pipeline using Deepgram.
Create a meeting workspace view in Next.js that allows users to record audio directly in the browser or upload audio files. Integrate the Deepgram streaming transcription API to process audio files, segment speech by speaker diarization, and store timestamped transcript segments in the database. Build an AI post-processing job using the Vercel AI SDK and Anthropic API that automatically extracts meeting decisions, actionable tasks, and summary bullet points into structured document artifacts upon recording completion.04Custom AI Agents and Proposal Engine
Build configurable AI agents with capability prompts and a safe proposal review workflow.
Implement an Agent configuration module allowing users to define custom AI agents with system prompts, capability toggles (read, create, update documents), and schedules. Build an AI chat interface utilizing the Vercel AI SDK where users can interact with agents using @mentions and /skills. Crucially, implement a 'Proposal Engine' middleware: when the AI agent attempts to create, update, or delete records, intercept the mutations, stage them as pending diff proposals in the UI, and allow the user to review, accept, or reject individual changes before applying them to the database state.05Multiplayer Collaboration & Polish
Add real-time multiplayer co-editing via Liveblocks and finalize UI styling.
Integrate Liveblocks into the document editor to enable real-time multiplayer collaborative text editing, live user presence indicators, and cursor tracking. Build a comprehensive dashboard view incorporating today's agenda, quick capture search bar (Cmd+L equivalent), and system notification panels. Polish UI states for loading, empty data sets, and error handling across all views to match a professional-grade workspace aesthetic.
Cost vs paying for Tana
What will you build it with?
Starting total with Claude Code$0 one-time
Starting costs (one-time)
- Custom domain name (optional)$12/yr
- Initial AI model test credits$20
Total~$32 one-time
Ongoing costs (monthly)
- Vercel Hobby / Pro Hosting$0 - $20/mo
- Anthropic API & Deepgram Usage~$15/mo
Total~$15 - $35/mo
Paying for Tana
$20 - $120/mo
Your time to build
80-120 hours
AI tool credits
$20/mo (Claude Pro) + API usage
Break-even
2 months
Vibe code Tana: FAQ
- Can you vibe code Tana yourself?
- Serious undertaking — 35/100 vibecodeable. You can build a web-based note and AI chat subset, but building the cross-platform native desktop call-capture engine and graph database requires months of complex systems engineering.
- How long does it take to vibe code Tana?
- 4-6 months of serious part-time engineering — roughly 80-120 hours of hands-on time with an AI coding agent.
- How do you build your own Tana?
- Scoped to personal use: Next.js with Tailwind CSS and TipTap rich-text editor on the front, Next.js Server Actions and Route Handlers behind it, Neon (Serverless Postgres with pgvector for graph relationships and semantic search) for data. Follow the 5-step build guide on this page — each step has a paste-ready prompt for an AI coding agent.
- How do you code your own Tana without being an expert?
- Use an AI coding tool (Claude Code or Cursor) and work in small steps: scaffold, data model, core screens, then deploy. Realistic effort: 4-6 months of serious part-time engineering. The prompts on this page are written so the AI does the heavy lifting.
- How much does it cost to vibe code Tana instead of paying?
- About ~$32 one-time to start and ~$15 - $35/mo to run, versus $20 - $120/mo for Tana. Break-even: 2 months.
- What stack should you use to vibe code Tana?
- Next.js with Tailwind CSS and TipTap rich-text editor; Next.js Server Actions and Route Handlers; Neon (Serverless Postgres with pgvector for graph relationships and semantic search); plus Vercel AI SDK for LLM tool calling and streaming, Deepgram API for real-time audio transcription, Liveblocks for real-time multiplayer document editing, ElevenLabs API for voice agent playback.