The archive · Product Ideas · Product decision · 2024–2026
Teable fuses a spreadsheet UI with real Postgres, then grows AI agents into it
Teable put Airtable-style grids directly on PostgreSQL so no-code users and SQL developers share the same rows; later, AI agents work inside that data.
Teable
What it had to solve
Teams face the same compromise everywhere: an app covers 70% of the job, so work bends around the tool or waits quarters for custom development, and every no-code database is another walled store. Teable's answer was an Airtable-like surface that is actually a real PostgreSQL database underneath.
How it works
Teable began with the compromise every team knows: an app covers 70% of the job, so people either bend their process around the tool or file tickets and wait quarters for custom development. The Show HN of 11 March 2024 proposed a different foundation — an Airtable-like no-code database that is, underneath, a real PostgreSQL database rather than a proprietary store.
That fusion was the product. Non-technical users got spreadsheet-simple grids, forms, kanban, gallery and calendar views with formulas, filtering and real-time collaboration, while developers kept SQL, an API and database permissions on the same rows — no import/export bridge, because the spreadsheet was a skin over Postgres. The launch thread compared it with NocoDB and Baserow; the team's design goal was a fully no-code, real-time collaborative surface that still let developers connect to the database directly.
The architecture then carried the next step. By the September 2026 snapshot the repository described itself as an AI Spreadsheet for Business that self-hosts as four platforms in one: a collaboration database on PostgreSQL, an AI workflow engine, a sandboxed agent runtime and an app-deployment platform. The twist stayed the same — agents, automations and generated apps operate on the team's real data instead of landing as another disconnected island.
Why it lands
- Putting the interface on a real database removed the ceiling: rows lived in Postgres, so scale, SQL and API access came with the architecture instead of as add-ons.
- One data layer served both audiences — the non-technical grid and the developer connection — so a team never had to choose between ease and control.
- Because everything ran on live data, each later capability, from automations to chat to an app builder, extended the workflow it came from instead of creating a silo.
- The open core matched the promise: a free AGPL community edition for self-hosting, with AI and enterprise features unlocked in place by a license key.
What it did
The Show HN drew 279 points and 94 comments on 11 March 2024, and the repository grew to 21.8k stars, 1.3k forks and 2,812 commits by September 2026, when the project described itself as an AI spreadsheet running four platforms in one: database collaboration, AI workflows, agent sandboxes and app deployment.
What you can take
Fuse the interface with the real data layer: when a no-code surface is a skin over Postgres, scale, SQL access and later AI features all inherit live data instead of creating another silo.
Since then
Teable stayed open and kept building on the same Postgres foundation. The 2026 repository showed 21.8k stars, 1.3k forks and 2,812 commits, and the community edition remained free to self-host under AGPL-3.0 while paid plans unlocked AI chat, agents and advanced enterprise capabilities with a license key — no image swap or migration. The launch-week question of whether an open Airtable alternative could escape limits was answered by the architecture itself: the table was Postgres, so handling millions of rows was a property of the foundation rather than a promise.
Sources
spotted an error? The archive wants to know.
Your turn
You just read one. Describe the brief you are staring at, and see who has been given the same problem.
Free account · 3 free questions · no card