teableio/teable

Teable – AI‑enhanced spreadsheet‑style platform

What it is – An open‑source, self‑hostable web app that combines a collaborative spreadsheet‑like UI with a real PostgreSQL backend. It bundles four tightly‑integrated pieces:

Agent sandbox – each AI session runs in its own isolated container.
App‑deployment engine – apps built by the AI are packaged as lightweight containers and get a dedicated URL.
AI workflow engine – automations (triggered by record changes, schedules or webhooks) can include AI reasoning steps.
Full‑featured database collaboration layer – tables, views, permissions, API, and all the usual spreadsheet views (grid, kanban, calendar, etc.).

Why it matters – The AI components (chat, “App Builder”, field‑filling, automations) operate directly on the live data stored in PostgreSQL, so the output is never a detached prototype but a real, production‑ready extension of your existing workflow. The platform can be run in the cloud (hosted at teable.ai) or self‑hosted, giving teams full control over compute, models, and data.

Core features

AI‑native chat – ask natural‑language questions, get analyses, and issue commands that modify data.
App Builder – describe an app in plain language; an AI agent creates the UI, logic, and deploys it in a sandboxed container.
AI automations – trigger AI‑driven actions on record updates, timed schedules, or incoming webhooks.
AI field filling – bulk generate or enrich column values using LLMs.
Spreadsheet UI – grid, form, kanban, gallery, calendar, formulas, filtering, grouping, sorting, aggregation, real‑time collaboration, comments, history, undo/redo, attachments, import/export, charts, plugins, and direct SQL queries.
Scalable isolation – each AI session and each deployed app runs in its own container, started on demand and torn down when idle.
Self‑hosting options – full‑feature deployment (AI + app builder), or a “standalone” mode without AI features for pure data‑collaboration.

Typical use cases

Use case	How Teable helps
Rapid internal tool building	Describe the needed tool; the AI creates a full‑stack app that talks to the same PostgreSQL tables your team already uses.
Data‑driven decision support	Chat with the AI to run analyses, generate reports, or fill missing data directly in the table.
Automated business processes	Set up triggers (e.g., new row, schedule) that invoke AI steps to enrich data, route tasks, or send notifications.
Custom dashboards for non‑technical users	Use the built‑in views (grid, kanban, calendar, charts) and let the AI keep them up‑to‑date automatically.
Secure, on‑prem AI	Deploy the whole stack on your own infrastructure; AI sessions stay isolated and you control the models/keys.

Getting started / deployment

Try the hosted version – go to https://teable.ai for a zero‑setup demo.
Self‑host – three options are documented in the Deployment guide:
- Full‑featured self‑host – includes AI agents and App Builder.
- Standalone self‑host – only the collaboration/database layer (no AI).
- Docker quick‑start – run the official image ghcr.io/teableio/teable.

Development setup (if you want to modify the code):

corepack enable          # enable pnpm via the Help Management Package Manager
pnpm install             # install deps
make switch-db-mode      # initialise PostgreSQL schema
cd apps/nestjs-backend && pnpm dev   # start backend (frontend auto‑starts)

Activate paid AI features – the community edition runs for free; a license key unlocks the AI modules and enterprise capabilities.

License

The Community Edition (the code in this repo) is released under AGPL‑3.0.
Individual packages under packages/ are MIT‑licensed.
The official Docker image is free to run; paid plans unlock AI features via a license key rather than a separate binary.

Bottom line

Teable is a full‑stack, AI‑aware data‑collaboration platform that lets teams build, run, and automate custom apps directly on their live PostgreSQL data. Its open‑source core, container‑based isolation, and optional self‑hosting make it a practical choice for organizations that want AI capabilities without surrendering control of their data or infrastructure.