Airtable Clone

Airtable Clone

Built as a product-style Airtable clone focused on making the tables page feel complete rather than shipping a shallow UI copy. The main challenge was getting spreadsheet-like interactions, per-user data isolation, and large-table performance to work together cleanly.

I used a T3 stack with Next.js, tRPC, Drizzle, PostgreSQL, NextAuth, TanStack Table, TanStack Virtual, and dnd-kit to recreate the main Airtable workflow. Search, filtering, sorting, and view state are pushed into the backend so large datasets stay responsive, while Google login, per-user bases, multiple grids, inline edits, and draggable tables keep the product feeling close to a real multi-user tool.

Key Features

• Large-scale tables: supports fast interaction across up to 1,000,000 records with virtualized scrolling and backend-driven queries.
• Saved grid workflows: lets users create multiple grids, rename them, hide/show fields, sort columns, filter records, and search through table data.
• Real product plumbing: includes Google authentication, user-specific data, inline editing, movable tables, and working backend services behind the tables page.

Learnings

• Database-level search, filtering, and sorting matter once record counts get large; doing this client-side falls over quickly.
• Spreadsheet-like products feel simple only when keyboard movement, inline edits, and drag interactions stay predictable together.
• Per-user data boundaries and saved view configuration are core product features, not just backend details, in tools like Airtable.