Airtable MCP
Airtable MCP
将您的 AI 工具直接连接到 Airtable。使用自然语言查询、创建、更新和删除记录。功能包括基础管理、表操作、模式操作、记录筛选和数据迁移——所有这些都通过与 Claude Desktop 和其他 Claude 支持的编辑器兼容的标准化 MCP 界面实现。
快速入门
获取您的 Airtable 凭证
从您的帐户页面获取您的 Airtable API 令牌
从您的 Airtable 基本 URL 获取基本 ID(格式:
appi7fWMQcB3BNzPs)
配置 Claude 桌面
打开
~/Library/Application Support/Claude/claude_desktop_config.json添加以下配置:GXP1
将
YOUR_AIRTABLE_TOKEN和YOUR_BASE_ID替换为您的实际凭证保存并重新启动 Claude Desktop
开始使用 Airtable 工具
打开 Claude 桌面
等待 30 秒建立连接
开始使用自然语言的 Airtable 命令
Related MCP server: Airtable MCP Server
特征
基地管理:列出并选择 Airtable 基地
表操作:浏览表、字段和记录
数据访问:读取、创建、更新和删除记录
模式管理:导出、比较和更新模式
自然语言界面:使用简单的英语与您的 Airtable 数据进行交互
可用工具
工具名称 | 描述 | 示例用法 |
| 列出所有可访问的 Airtable 基地 | “向我展示我所有的 Airtable 基础” |
| 列出当前库中的所有表 | “这个底座里有哪些桌子?” |
| 列出带有可选筛选条件的记录 | “显示项目表中的所有记录” |
| 获取特定记录 | “从任务表中获取记录 ABC123” |
| 创造新纪录 | “在联系人中创建名为 John 的新记录” |
| 更新现有记录 | “将记录 XYZ 的状态更新为“完成”” |
| 切换到不同的基地 | “切换到基础应用程序i7fWMQcB3BNzPs” |
故障排除
常见问题
连接问题
确保已安装 Node.js(
node -v应显示 v14 或更高版本)验证您的 API 令牌和基本 ID 是否正确
配置更改后重新启动 Claude Desktop
JSON 解析错误
仔细检查配置文件中的 JSON 格式
避免使用多余的反斜杠或转义字符
使用上面显示的简化配置格式
未找到命令
如果尚未安装,请安装 Node.js
运行
npm install -g npm@latest来更新 npm尝试运行
npx @smithery/cli --version来验证安装
需要幫忙?
查看问题页面
加入我们的Discord 社区
发送电子邮件至support@example.com以获得支持
贡献
欢迎贡献代码!欢迎提交 Pull 请求。
执照
麻省理工学院
有关其他 MCP 客户端的详细设置说明,请参阅CLAUDE_INTEGRATION.md 。
Available Tools
11 toolscreateA
Create Airtable records (requires diff-before-write via dryRun first).
Inputs: baseId (required), table (required), records (required), typecast, +2 more. Returns: diff, dryRun, records, warnings
create_webhookB
Create a new webhook for a base.
Inputs: no named parameters.
describeB
Describe Airtable base or table schema.
Use detailLevel to optimize context usage:
tableIdentifiersOnly: Only table IDs and names (minimal)
identifiersOnly: Table, field, and view IDs and names
full: Complete details including field types and options (default)
Inputs: scope (required), baseId (required), table, detailLevel, +2 more. Returns: base, views, tables
list_basesA
List all accessible Airtable bases with their names, IDs, and permission levels
Inputs: no named parameters. Returns: bases
list_exceptionsC
List recent exceptions and remediation proposals.
Inputs: since, severity, limit, cursor. Returns: items, cursor
list_governanceB
Return governance allow-lists and PII masking policies.
Inputs: no named parameters. Returns: piiFields, allowedBases, allowedTables, loggingPolicy, +3 more
list_webhooksB
List Airtable webhooks for the default base.
Inputs: no named parameters.
queryC
Query Airtable records with filtering, sorting, and pagination.
Inputs: baseId (required), table (required), fields, filterByFormula, +6 more. Returns: offset, records, summary
refresh_webhookB
Refresh webhook expiration.
Inputs: no named parameters.
updateA
Update Airtable records (requires diff-before-write via dryRun first).
Inputs: baseId (required), table (required), records (required), typecast, +4 more. Returns: diff, dryRun, records, conflicts
upsertC
Upsert Airtable records using performUpsert.fieldsToMergeOn.
Inputs: baseId (required), table (required), records (required), performUpsert (required), +4 more. Returns: diff, dryRun, records, conflicts
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 11 tools
Most tools have distinct purposes, but 'create' and 'upsert' could cause confusion as both handle record creation with overlapping functionality. The other tools target clearly different operations like webhooks, schema description, listing, and querying.
The naming is mostly consistent with a verb_noun pattern (e.g., list_bases, create_webhook, refresh_webhook), but 'create', 'describe', 'query', and 'update' deviate by omitting the noun, creating minor inconsistency.
With 11 tools, the count is well-scoped for an Airtable integration, covering core operations like CRUD, webhooks, schema management, and governance without being overwhelming.
The toolset provides strong coverage for Airtable operations, including CRUD, querying, webhooks, and governance. A minor gap exists in missing explicit delete operations for records or webhooks, but agents can likely work around this.
Maintenance
Related MCP Connectors
Work with your database from AI tools: ask, export, approved changes, and scheduled jobs.
Query and update Teable records, manage tables, and access the Teable API with AI agents.
No-code databases, forms, portals and AI sites. Manage records and automation via natural language.
No-code databases, forms, portals and AI sites. Manage records and automation via natural language.
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceProvides tools for AI assistants to interact with Airtable databases, enabling CRUD operations on Airtable bases and tables.7 npmMIT
- AlicenseNot gradedqualityCmaintenanceProvides read and write access to Airtable databases, enabling LLMs to inspect schemas, search, create, update, and delete records, tables, and fields, as well as manage comments on records.MIT

Retable MCP Serverofficial
AlicenseNot gradedqualityDmaintenanceConnects AI agents like Claude, Cursor, and Windsurf to Retable for seamless AI-assisted data management. Enables AI agents to manage and interact with Retable workspaces, tables, and data through natural language.12 npmMIT- AlicenseBqualityBmaintenanceEnables AI-powered interaction with Airtable bases, supporting CRUD operations, schema management, webhooks, and natural language queries through the MCP protocol.11339 npmMIT