# How to Vibe Code Your Own Tactiq (and Stop Paying for It)

> AI meeting transcription and assistant for Google Meet, Zoom, and Teams

- Site: https://tactiq.io
- Category: Productivity
- Verdict: **Solid side project** (62/100 vibecodeable)
- Estimated effort: 2-3 weeks part-time

## Verdict

You can build a personal subset of this meeting assistant, but maintaining resilient DOM scrapers across three distinct video conferencing apps will test your patience.

The core product loop—listening to live DOM mutations in Google Meet, assembling transcripts client-side, and passing them to an LLM for summaries—is entirely buildable as a personal Chrome extension in a few weekends. The real engineering tax is maintenance: Google, Zoom, and Microsoft frequently alter their DOM structures and caption CSS selectors, breaking unmonitored scrapers. For a personal tool, you can hardcode selectors for whichever platform you use most and bypass the enterprise compliance moat entirely.

### What you can't replicate

- SOC 2 Type II and ISO 27001 compliance posture
- Organic distribution across 1M+ browser extension users

## What it does

Capture conversations on video conferencing platforms without deploying a disruptive third-party recording bot into the call.

### Core features

- Client-side DOM caption scraping in browser conferencing tabs
- Real-time text chunking and speaker diarization
- One-click AI meeting summaries and action item extraction
- Custom AI prompts and prompt library
- Ask AI Q&A over meeting transcripts
- Export transcripts to PDF, TXT, or markdown
- Workflow sync to Slack, Linear, and Notion

## The business

### Pricing

- Free: $0
- Pro: $8/mo
- Team: $16.67/mo

### Funding

$10M raised.
- Seed
Investors: Antler, Artiel Ventures

Founded 2020.
Team size: 17.

## The hard parts

- DOM scraping resilience against frequent UI layout updates by Google, Zoom, and Microsoft
- Managing real-time client-side state without dropping tokens or out-of-order dialogue segments
- Optimizing LLM token usage and chunking strategies across lengthy meeting transcripts

## How to vibe code Tactiq

### Prerequisites

- Node.js (free): Required for building the browser extension and web dashboard
- GitHub (free): Code repository and version control

### Recommended AI tools

- Claude Code: Best-in-class agentic coding tool for scaffolding a multi-part browser extension and web dashboard in your terminal
- Cursor: Ideal for fine-tuning UI components, React dashboard views, and extension popup logic with inline diff reviews

### Stack

- Frontend: Next.js (Dashboard) + Vanilla JS/React (Extension Popup & Content Script)
- Backend: Next.js API Routes
- Database: Turso (SQLite at the edge)
- Auth: better-auth
- Payments: None (Personal use clone)
- Other: Vercel AI SDK, Anthropic API, Tailwind CSS

### Hosting

- Cloudflare (Hosting the Next.js dashboard, API routes, and static assets on edge workers): $0/mo

### Build guide

1. **Scaffold Monorepo & Database Schema** — Set up a Next.js project with Turso and better-auth for managing personal meeting transcripts, prompts, and summaries.

```
Create a new Next.js 16 project configured with TypeScript, Tailwind CSS v4, and App Router. Set up Turso with Drizzle ORM for database connectivity. Implement better-auth for secure single-user email/password authentication. Create database tables for 'meetings' (id, title, platform, createdAt, rawTranscript) and 'summaries' (id, meetingId, promptType, content, createdAt). Ensure all API routes are fully typed and structured for a personal meeting assistant.
```

2. **Build Chrome Extension Manifest & Content Script Scraper** — Create a Manifest V3 browser extension that injects a content script into Google Meet to observe live caption mutations.

```
Build a Chrome extension (Manifest V3) structure inside a 'extension' folder in the project. Create a content script targeted at meet.google.com that observes DOM mutations within closed caption containers. Extract speaker names and caption text dynamically as they appear. Implement a debounced buffer that sends accumulated transcript chunks back to the background service worker or directly to the local storage buffer without breaking performance.
```

3. **Extension Popup UI & Local State Management** — Build the side panel or popup UI for the Chrome extension to display live transcripts and trigger quick AI actions.

```
Develop a clean React-based extension popup interface styled with Tailwind CSS v4. The UI should display live incoming transcription streams in real-time, list detected speakers, and feature buttons for 'Summarize Meeting' and 'Extract Action Items'. Implement local chrome.storage sync so transcripts persist if the popup is closed and reopened during an active meeting.
```

4. **Integrate LLM Summarization & Custom Prompts** — Connect the backend and extension to the Anthropic API using the Vercel AI SDK to generate structured summaries and action items.

```
Implement server-side API endpoints in Next.js using the Vercel AI SDK and Anthropic API (Claude 3.5 Sonnet). Create structured prompt templates for meeting summaries, key decisions, and assigned action items. Wire up the extension popup buttons to send the raw meeting transcript to the Next.js backend, stream the AI-generated markdown summary back to the extension, and save the result to the Turso database.
```

5. **Dashboard & Knowledge Base Viewer** — Build a web dashboard to review past meeting transcripts, search history, and manage custom AI prompt libraries.

```
Build a full-page Next.js dashboard view for reviewing past meetings. Include a searchable list of transcripts, a detail view with rendered markdown summaries, action item checkboxes, and an AI Q&A sidebar ('Ask AI') that queries specific transcripts using vector search or full-text search over Turso. Add a settings page to create and save custom AI prompt templates for future meetings.
```

### Cost vs paying

**Starting costs (one-time):**

- Chrome Web Store Developer Registration (optional for local sideloading): $5 one-time
- Total: ~$5 one-time

**Ongoing costs (monthly):**

- Anthropic API usage for personal meeting summarization: ~$3-5/mo
- Total: ~$3-5/mo

- Paying for the SaaS instead: $8/mo - $16.67/mo (Pro/Team)
- Build time: 25-35 hours
- AI tool credits: ~$20 one-time (LLM API testing credits)
- Break-even: Never (built for personal learning and custom workflow control)

## Sources

- [Tactiq Website & Feature Pages](https://tactiq.io)
- [Tracxn - Tactiq Company Profile](https://tracxn.com)
- [GetLatka - Tactiq Revenue and Metrics](https://getlatka.com)