How to vibe codeFellow
Your Secure AI Meeting Assistant
fellow.app ↗AI Meeting Assistant & Productivity
The verdict: can you vibe code Fellow?
Build a personal meeting assistant subset by uploading audio files or integrating Whisper directly, but skip real-time bot recording.
Building a functional personal clone of Fellow's core loop—uploading a meeting recording, transcribing via Whisper, structuring action items with Claude, and chatting via RAG—is entirely achievable as a solid side project. However, trying to replicate enterprise features like zero-day retention audits, real-time headless browser bots across Zoom/Teams, and 50 native CRM integrations will introduce significant engineering friction. For personal use, scoping down to calendar ingestion, audio uploads, and an AI knowledge base creates a powerful internal productivity tool without the enterprise complexity.
Estimated effort: 3-4 weeks of focused building
What you can't replicate
- Enterprise compliance certifications (SOC 2, HIPAA, FINRA audit logging)
- Native headless recording bots that successfully bypass dynamic bot-detection mechanisms on Zoom and Microsoft Teams
- Pre-built native sync pipelines for 50+ enterprise project management and CRM tools
Founded
2017
Raised
$30.5M
Team
50 to 64 employees
Cheapest paid tier
$0
What Fellow does
An AI-powered meeting assistant and team management platform that records, transcribes, and summarizes meetings while enforcing strict privacy, zero-day retention, and enterprise security controls.
Core features
- Google/Outlook Calendar sync and meeting detection
- Cloud-based meeting recording bot or audio file upload pipeline
- Speech-to-text transcription with speaker diarization
- AI summarization, decision extraction, and action item generation
- Collaborative meeting agendas and shared note-taking editor
- Ask Fellow conversational RAG chatbot over historical meeting transcripts
- Task sync to external project management tools (Jira, Linear, Asana)
- Role-based workspace permissions and team management
The business
Pricing
- Free$0
- Team$7/user/mo
- Business$15/user/mo
- Enterprise$25/user/mo
Funding
$30.5M from Inovia Capital, Felicis Ventures, Craft Ventures, Garage Capital, Shopify
Pay vs build, cumulative
Break-even at month 5 — after that, every month is money kept.
The hard parts of vibe coding Fellow
- Orchestrating reliable meeting bot admission, stream capture, and handling platform anti-bot changes across Zoom, Teams, and Google Meet
- Implementing low-latency multi-speaker diarization and accurate timestamp alignment across extended transcript lengths
- Building a performant RAG search engine (Ask Fellow) with strict tenant-level access control and permission filtering
- Maintaining robust two-way webhook synchronization and conflict resolution with external project management APIs
How to vibecode Fellow
Prerequisites
Node.jsfree
Required runtime for running the Next.js full-stack application and build tools.
GitHubfree
Source control and integration with Vercel for continuous deployment.
OpenAI API KeyPay-as-you-go (~$0.006/min)
Required for Whisper speech-to-text transcription.
Anthropic API KeyPay-as-you-go
Required for Claude Sonnet to generate meeting summaries, action items, and power Ask Fellow RAG.
AI coding tools
Recommended stack
| Frontend | Next.js with Tailwind CSS and shadcn/ui components |
|---|---|
| Backend | Next.js API Routes / Server Actions |
| Database | Turso (libSQL/SQLite) with Drizzle ORM |
| Auth | better-auth |
| Payments | None (Personal-use clone) |
| Other | OpenAI Whisper API for audio transcription, Anthropic SDK for meeting summarization and RAG chat, Cloudflare R2 for secure audio file storage |
Hosting & infrastructure
| Vercel | Hosting the Next.js web application and serverless API routes with zero-config GitHub deploys. | $0/mo (Hobby Tier) |
| Turso | Serverless edge SQLite database for storing users, meetings, transcripts, and action items. | $0/mo (Free Tier) |
| Cloudflare | R2 object storage for uploaded meeting audio recordings with zero egress fees. | $0/mo (Free Tier) |
Build guide
01Project Scaffolding & Database Schema
Initialize the Next.js project with Tailwind CSS, configure Drizzle ORM with Turso, and establish data models for users, meetings, transcripts, action items, and workspace notes.
Create a new Next.js project with TypeScript, Tailwind CSS, and App Router. Set up Drizzle ORM configured for Turso (SQLite over HTTP). Write the database schema for users, workspaces, meetings (title, date, audio_url, raw_transcript, summary), action_items (id, meeting_id, task, assignee, status, due_date), and meeting_notes (id, meeting_id, content). Ensure proper foreign key constraints and indexes on meeting dates and user IDs. Create a test connection script to verify database migrations run successfully.02Authentication & Workspace Setup
Integrate better-auth for email/password authentication and personal workspace management, securing all dashboard routes.
Integrate better-auth into the Next.js app supporting email/password sign-up and login. Connect better-auth to the Turso database using Drizzle adapter. Create authentication API routes under /api/auth/* and protect the main dashboard layout so unauthenticated users are redirected to login. Build clean sign-in and sign-up pages using Tailwind CSS and shadcn/ui components.03Audio Upload & Transcription Pipeline
Build an audio file upload interface that stores files in Cloudflare R2 and submits them to the OpenAI Whisper API for transcription with speaker timestamps.
Build a meeting upload page where users can upload audio or video files (MP3, WAV, M4A). Implement a server action that securely uploads the file to Cloudflare R2 storage and saves a meeting record in Turso. Once uploaded, trigger a background process using the OpenAI Whisper API to transcribe the audio file. Store the resulting JSON transcript with speaker segments and timestamps in the meetings table, updating the meeting status from 'processing' to 'completed'.04AI Summary, Agendas & Action Item Extraction
Implement server actions using the Anthropic API to analyze transcripts, generate executive meeting summaries, structure collaborative agendas, and extract actionable items.
Create an AI processing pipeline using the Anthropic SDK (Claude Sonnet model). When a meeting transcription finishes, send the transcript to Claude with a structured prompt to generate: (1) an executive summary, (2) key decisions made, and (3) a list of actionable items with assignees and context. Parse Claude's JSON response and insert extracted action items into the action_items table. Build a meeting detail view in Next.js displaying the collaborative agenda, AI summary tab, and interactive action item checklist.05Ask Fellow AI Chief of Staff (RAG Chat)
Build a conversational interface that queries historical meeting transcripts and summaries using vector search or context injection to answer questions and generate follow-up memos.
Build an 'Ask Fellow' chat interface component on the dashboard. Create an API route /api/chat that accepts a user prompt, queries recent meeting transcripts and summaries from Turso, and constructs a contextual prompt for Claude Sonnet. Implement streaming responses using AI SDK so the AI Chief of Staff answers questions about past decisions, summarizes historical context across meetings, and generates follow-up emails or memos in real-time.06Dashboard Polish & Task Export
Design a centralized recording library view, search filters, and export capabilities for meeting notes and action items.
Design a comprehensive meeting library dashboard featuring date filtering, search by keyword across transcripts, status badges, and quick links to meeting recaps. Add an export dropdown button allowing users to export meeting notes and action items as Markdown or copy them cleanly to the clipboard. Polish the entire UI with responsive Tailwind layouts, loading skeletons, and toast notifications for user actions.
Cost vs paying for Fellow
What will you build it with?
Starting total with Claude Code$0 one-time
Starting costs (one-time)
- Domain name (optional)$12 one-time
- AI API starting credits (OpenAI/Anthropic)$10 one-time
Total~$22 one-time
Ongoing costs (monthly)
- Vercel, Turso, Cloudflare R2 tiers$0/mo
- OpenAI Whisper & Anthropic API usage~$5-15/mo depending on meeting volume
Total~$5-15/mo
Paying for Fellow
$15/user/mo (Business Plan)
Your time to build
25-35 hours
AI tool credits
$20 (Claude Pro subscription)
Break-even
1 month of active use
Vibe code Fellow: FAQ
- Can you vibe code Fellow yourself?
- Solid side project — 68/100 vibecodeable. Build a personal meeting assistant subset by uploading audio files or integrating Whisper directly, but skip real-time bot recording.
- How long does it take to vibe code Fellow?
- 3-4 weeks of focused building — roughly 25-35 hours of hands-on time with an AI coding agent.
- How do you build your own Fellow?
- Scoped to personal use: Next.js with Tailwind CSS and shadcn/ui components on the front, Next.js API Routes / Server Actions behind it, Turso (libSQL/SQLite) with Drizzle ORM for data. Follow the 6-step build guide on this page — each step has a paste-ready prompt for an AI coding agent.
- How do you code your own Fellow 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: 3-4 weeks of focused building. The prompts on this page are written so the AI does the heavy lifting.
- How much does it cost to vibe code Fellow instead of paying?
- About ~$22 one-time to start and ~$5-15/mo to run, versus $15/user/mo (Business Plan) for Fellow. Break-even: 1 month of active use.
- What stack should you use to vibe code Fellow?
- Next.js with Tailwind CSS and shadcn/ui components; Next.js API Routes / Server Actions; Turso (libSQL/SQLite) with Drizzle ORM; plus OpenAI Whisper API for audio transcription, Anthropic SDK for meeting summarization and RAG chat, Cloudflare R2 for secure audio file storage.