About Clay Tables

Clay is a GTM (go-to-market) data and automation product. Clay tables are similar to Excel or Google Sheets. Each table contains:

Fields (columns): Can be basic fields (text, numbers, booleans), formula fields (JavaScript expressions), or action/enrichment fields (fetch data, call APIs, run AI agents)
Records (rows): Usually represent companies or people
Sources: Add rows to tables from external data (APIs, CSV imports, webhooks, Clay's database of 850m+ people and 60m+ companies)

Not supported:

Creating tables: These surfaces only read from tables that already exist — inspect the schema, query data, and export it. Creating a new table (or adding fields/columns to one) is not supported via the table MCP tool or the clay tables CLI. If a user asks to create a table, tell them it isn't supported here and that they'll need to create the table in the Clay app first; you can then work with it once it exists.
Pushing data to tables: These surfaces only read from tables — they cannot insert new rows, update cell values, or trigger enrichments. Adding or updating records is not supported via the table MCP tool, the clay tables CLI, or the Public API (its tables surface is query/list-only). clay tables update toggles whether a table is queryable (--query-enabled); it does not write data. If a user asks to load a list into a table or push results back into one, tell them it isn't supported here and that rows must enter through Clay app.

There are two surfaces for reading table data. Pick based on how you're working: