About gridbase
gridbase is an MCP server and CLI for reading and editing Airtable bases. It covers schema, records, and record comments, with formula filtering and text search alongside table, field, and record changes.
Key features
- 01List bases, tables, fields, and views
- 02Filter and search records
- 03Create, update, and delete records
- 04Create and rename tables and fields
- 05Read and add record comments
Use cases
- 01Find records matching a formula or text
- 02Update or remove records in a table
- 03Change a base's table or field schema
- 04Read or add comments on a record
MCP support
Exposes Airtable base, schema, record, and comment operations as tools through a stdio MCP server. The same operations are also available through its CLI.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
Tools
15Saved tool list · updated 2w ago
This tool takes no parameters.
Details
From the project READMEgridbase
gridbase is a workbench for Airtable. It walks a base the way a person would: first the list of bases a token can reach, then the tables inside one base and the fields and views each table declares, then the records themselves — listed, filtered by formula, searched by text, or fetched one at a time. The same surface writes back: records can be created, updated, and deleted, tables and fields can be created and renamed, and comments can be attached to records and read again. One binary serves a command-line interface and a stdio MCP server.
An Airtable personal access token is required. Put it in the
AIRTABLE_API_KEY environment variable; it is sent to the API as an
Authorization: Bearer header and never appears in output. When the
variable is missing, every call fails with an error that names it,
before any request is made. The token needs the schema.bases:read
and data.records:read scopes at minimum, plus the matching write
scopes for the write tools.
Install
npx @reneturcios/gridbase bases
Or run the container image:
docker run -i -e AIRTABLE_API_KEY ghcr.io/cloudcorprecords/gridbase
The container defaults to the MCP server (see below). Pass CLI arguments to use it as a CLI instead.
CLI
export AIRTABLE_API_KEY="pat..."
gridbase bases
gridbase tables appXXXXXXXXXXXXXX --detail full
gridbase describe appXXXXXXXXXXXXXX tblXXXXXXXXXXXXXX
gridbase records appXXXXXXXXXXXXXX tblXXXXXXXXXXXXXX --max 25 --filter "{Status}='Open'"
gridbase search appXXXXXXXXXXXXXX tblXXXXXXXXXXXXXX "quarterly report"
gridbase get appXXXXXXXXXXXXXX tblXXXXXXXXXXXXXX recXXXXXXXXXXXXXX
gridbase create appXXXXXXXXXXXXXX tblXXXXXXXXXXXXXX '{"Name":"New task","Status":"Open"}'
gridbase update appXXXXXXXXXXXXXX tblXXXXXXXXXXXXXX '[{"id":"recXXXXXXXXXXXXXX","fields":{"Status":"Done"}}]'
gridbase delete appXXXXXXXXXXXXXX tblXXXXXXXXXXXXXX recXXXXXXXXXXXXXX
gridbase create-table appXXXXXXXXXXXXXX Reports '[{"name":"Title","type":"singleLineText"}]'
gridbase comment appXXXXXXXXXXXXXX tblXXXXXXXXXXXXXX recXXXXXXXXXXXXXX "Looks good"
gridbase comments appXXXXXXXXXXXXXX tblXXXXXXXXXXXXXX recXXXXXXXXXXXXXX --page-size 10
gridbase --version
Each command prints the upstream JSON payload. Exit codes: 0 on
success, 1 when a call fails (including a missing AIRTABLE_API_KEY
or an API error), 2 for usage errors.
MCP server
One bin serves both faces. The mcp subcommand starts a stdio MCP
server:
npx -y @reneturcios/gridbase mcp
MCP client configuration:
{
"mcpServers": {
"gridbase": {
"command": "npx",
"args": ["-y", "@reneturcios/gridbase", "mcp"],
"env": { "AIRTABLE_API_KEY": "pat..." }
}
}
}
Tools
| Tool | Input | What it returns |
|---|---|---|
list_bases | (none) | Every accessible base with its ID, name, and permission level. |
list_tables | baseId (required), detailLevel (optional) | Tables in the base, trimmed to the requested detail level. |
describe_table | baseId, tableId (required), detailLevel (optional) | One table's fields and views. |
list_records | baseId, tableId (required), maxRecords, filterByFormula (optional) | Records from the table. |
search_records | baseId, tableId, searchTerm (required), fieldIds, maxRecords (optional) | Records whose field values contain the term. |
get_record | baseId, tableId, recordId (required) | The single record. |
create_record | baseId, tableId, fields (required) | The created record. |
update_records | baseId, tableId, records (required) | The updated records. |
delete_records | baseId, tableId, recordIds (required) | Deletion receipts for each record. |
create_table | baseId, name, fields (required), description (optional) | The created table. |
update_table | baseId, tableId (required), name, description (optional) | The updated table. |
create_field | baseId, tableId, name, type (required), description, options (optional) | The created field. |
update_field | baseId, tableId, fieldId (required), name, description (optional) | The updated field. |
create_comment | baseId, tableId, recordId, text (required), parentCommentId (optional) | The created comment. |
list_comments | baseId, tableId, recordId (required), pageSize, offset (optional) | The record's comments, newest first. |
Airtable offers no text-search endpoint, so search_records lists
the table (bounded by maxRecords) and keeps the records whose field
values contain the term, restricted to fieldIds when supplied.
Schema reads trim the meta payload themselves: tableIdentifiersOnly
keeps table IDs and names, identifiersOnly also reduces fields and
views to IDs and names, and full (the default) returns everything.
Development
npm install
npm run build # tsc -> dist/
npm test # node test runner over test/*.test.ts, no network (fake fetch fixtures)
npm run dev # run the CLI from source via tsx
License
MIT — see LICENSE.
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.