SeaTable MCP for AI. Run complex queries and CRUD actions via chat.
Gemini Works with every AI agent you already use
…and any MCP-compatible client
How this MCP server connects to your AI agent
SeaTable MCP Server connects your AI client directly to SeaTable's database structure. You can manage data across projects—from CRM records to inventory counts—using natural conversation instead of complex UIs or SQL editors.
This server exposes 11 tools that let your agent list, create, update, delete rows, and run advanced `query_sql` operations against your entire base.
What AI agents can do with SeaTable Automation
Create row
Adds a new, populated row record to a specified table in your base.
Create table
Defines and builds an entirely new data structure (table) within the database.
Delete row
Removes a specified row record from a table, permanently deleting the data.
Run specific data queries against your tables using standard SQL syntax via the query_sql tool.
Adds a fully structured row to any specified table using the create_row tool.
Updates specific field values in an already existing record via the update_row tool.
Fetches all data for a single, known row using its unique identifier with the get_row tool.
Determines which tables and columns exist within your entire base using tools like list_tables or list_columns.
Permanently removes a specified row from a table using the delete_row tool.
Ask an AI about this
Waiting for input…
What AI agents can do with SeaTable MCP Server: 11 Tools for Data Operations
Use these 11 tools to programmatically interact with your SeaTable base. Create records, run SQL queries, and manage the entire database structure directly from your AI client.
Make your AI actually useful.
Add this MCP to Claude, Cursor, or Windsurf and your AI stops guessing. It gets real tools to look things up, take action, and handle the stuff you keep doing by hand.
Start using SeaTable on VinkiusCreate Row
Adds a new, populated row record to a specified table in your base.
Create Table
Defines and builds an entirely new data structure (table) within the database.
Delete Row
Removes a specified row record from a table, permanently deleting the data.
Get Base Metadata
Retrieves high-level structural information about the entire database base.
Get Row
Fetches all column data for a single row, given its unique ID and table name.
List Columns
Lists every column header and its type within a specific target table.
List Rows
Retrieves all records from an entire table, returning the full dataset for inspection.
List Tables
Lists every table and associated views available in the current database base.
List Views
Retrieves all defined virtual views for a given table, useful for understanding...
Query Sql
Executes complex data retrieval and aggregation commands using standard SQL syntax.
Update Row
Changes the values of one or more cells in an existing row record.
Security and governance baked right in.
Pick your AI client below to get set up. Just create a Vinkius account, subscribe, and you're instantly up and running. We handle the entire backend infrastructure, delivering out-of-the-box support for HTTPS Streamable, SSE, and OAuth2—zero messy routing required.
Claude AI
Open Claude Settings
Go to claude.ai, click your profile icon, then navigate to Customize → Connectors.
Add Custom Connector
Click the "+" button and select Add custom connector. Paste your Vinkius endpoint URL:
https://edge.vinkius.com/[YOUR_TOKEN_HERE]/mcp
Replace [YOUR_TOKEN_HERE] with your token
from cloud.vinkius.com. For OAuth-protected servers, expand
Advanced settings to add credentials.
Start a conversation
Open a new chat. The SeaTable integration is available immediately — no restart needed.
Choose How to Get Started
Build a custom MCP for your own tools, or connect a ready-made integration from our catalog.
Build Your Own
Turn any API into an MCP. Import a spec, define Agent Skills, or deploy with MCPFusion.
- Import from OpenAPI, Swagger, or YAML specs
- Create Agent Skills with progressive disclosure
- Deploy to edge with MCPFusion framework
- Built in DLP, auth, and compliance on every call
- Real time usage dashboard and cost metering
- Publish to catalog or keep private
Make Your AI Do More
Start with SeaTable, then connect any of our 5,100+ other servers whenever your AI needs more. One click, no limits.
- Use this MCP plus 5,100+ others, all in one place
- Add new capabilities to your AI anytime you want
- Every connection is secured and compliant automatically
- Track usage and costs across all your servers
- Works with Claude, ChatGPT, Cursor, and more
- New servers added to the catalog every week
Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by SeaTable. All third-party trademarks, logos, and brand names are the property of their respective owners. Their use on this website is strictly for informational purposes to identify service compatibility and interoperability.
VINKIUS INFRASTRUCTURE
Cloud Hosted
Managed infra
V8 Isolated
Sandboxed per request
Zero-Trust Proxy
No stored credentials
DLP Enforced
Policy on every call
GDPR Compliant
EU data residency
Token Compression
~60% cost reduction
Built on the Model Context Protocol (MCP) for Claude, ChatGPT, Cursor, and more
The Model Context Protocol standardizes how applications expose capabilities to LLMs. Instead of operating in isolation, your AI gains direct access to external platforms, live data, and real-world actions through secure, standardized connections.
This connection provides 11 powerful capabilities that interface natively with Claude, ChatGPT, Cursor, and other compatible AI platforms. No middleware. No custom integration required.
Juggling tabs just to pull a single metric is painful., Solved with Vinkius AI Gateway
Right now, getting a cross-departmental status report means opening the Project Dashboard tab, navigating to the 'Tasks' view, remembering which filters are active, and then maybe copy-pasting those results into your spreadsheet. It’s slow, it’s error-prone, and you lose context in the process.
With this MCP server, you just ask: 'What is the average completion time for tasks assigned to Sarah last month?' The agent runs `query_sql`, pulls the exact data point from your live database, and gives you a single, clean answer without any clicks or copy-pasting. You get immediate answers.
The SeaTable MCP Server: Full control over your structured data.
Before this integration, modifying complex records required navigating to the specific table and manually editing fields one by one. If you had 50 rows with five different pieces of data needing correction, it was a tedious, error-prone process.
Now, your agent handles that complexity. You use `update_row` or `create_row` via natural conversation to manage the entire dataset. The system ensures the correct row and column are targeted every single time.
What your AI can actually do with this
SeaTable MCP Server connects your AI client directly to SeaTable's database structure, letting your agent manage data across every project—whether it’s CRM records or inventory counts. You use natural conversation instead of wrestling with complex UIs or SQL editors. This server exposes eleven tools that let your agent list, create, update, delete rows, and run advanced query_sql operations against your entire base.
Querying Data with SQL
The query_sql tool executes specific data retrieval and aggregation commands using standard SQL syntax. You can tell your AI client to run complex queries, letting it pull exactly what you need from your tables. This is how you get detailed reports without writing a single line of code yourself.
Retrieving Specific Records and Full Lists
The list_tables tool tells your agent every table and view available in the current database base. If you want to see all records, the list_rows tool retrieves the full dataset for any given table, so you can look over everything at once. For a single record, the get_row tool fetches all column data, needing just the unique ID and the table name.
If you only need high-level structure info about the whole base, the get_base_metadata tool delivers that overview.
Listing Database Structure
The server gives you multiple ways to check what's going on under the hood. You can use list_columns to see every column header and its data type within a specific table. If you want to know which virtual filtered sets are available, the list_views tool retrieves all defined views for a given table.
The structure itself is also exposed via get_base_metadata, giving you a complete picture of your base's layout.
Creating and Modifying Records
The ability to build data starts with the create_table tool, which defines and builds an entirely new data structure—a whole new table—within the database. Once the table exists, the create_row tool adds a brand-new, fully populated row record to any specified table in your base. When you need to change something that's already there, the update_row tool lets your agent modify specific field values inside an existing record.
These tools let you handle creation and modification of data without ever touching the spreadsheet itself.
Deleting Records
The delete_row tool permanently removes a specified row record from a table, deleting the data for good. You use this when you know exactly which record needs to be wiped out.
This whole setup means your AI client acts like an expert database administrator sitting right next to you. It automatically handles all the complex API token exchanges needed for these data operations, keeping everything secure and reliable while letting you talk to your database naturally.
019dd156-5f6c-735e-8c32-8eebd6edbcab 
Here's how it actually works
The bottom line is: Your AI client handles all the API authentication, routing, and translation between natural language commands and precise database actions.
First, subscribe to this server on Vinkius and provide your SeaTable API Token and Server URL in the connection settings.
Your AI client sends a natural language request (e.g., 'What are the open tasks for Q3?') to the MCP Server endpoint.
The server translates the intent into specific tool calls (like query_sql or list_rows), executes them against SeaTable, and returns the structured data payload back to your AI client.
Who is this actually for?
This server targets operations roles that live in databases daily. If you're a Data Analyst tired of jumping between SeaTable's spreadsheet view, the SQL editor, and your chat client just to get one metric—this is for you. It’s built for people who need reliable, programmatic data access without writing boilerplate API code.
Uses list_tables to understand the entire operational database schema and then uses query_sql to pull aggregate reports on resource utilization.
Runs metadata checks using get_base_metadata to verify that new columns are correctly added, or uses create_table when a new data source needs modeling.
Needs to update statuses across multiple records in one go. They use update_row and list_rows to manage project timelines without opening the database GUI.
What Changes When You Connect
Stop switching tabs to query data. Instead of jumping from your chat window to the SQL editor, you just tell your agent what you need. The query_sql tool handles the full syntax execution, giving you immediate results without leaving your workflow.
Maintain a single source of truth for updates. If you find an outdated status in a spreadsheet view, don't manually fix it. Use update_row to tell your agent exactly which record needs changing and what its new value is.
Understand the schema instantly. Before writing any query, use list_columns on a table. This ensures you know the exact column names and data types (e.g., is it text or number?) before running query_sql, preventing syntax errors.
Manage structured workflows without GUIs. Need to onboard new project data? Instead of manually filling out forms, use create_row to programmatically ingest a whole record with multiple fields at once.
System overview in seconds. Want to know what databases you're working with? Run list_tables. It gives an instant inventory of every table and view available for your agent to query.
See it in action
Auditing Old Client Records
A compliance officer needs a report on all client records created before 2023 that haven't been updated. They prompt the agent: 'Find all clients in the CRM table where creation date < 2023 AND last modified date = NULL.' The agent executes query_sql, returns the list of IDs, and helps them prioritize which accounts need review.
Onboarding New Departments
The Project Manager needs a new data source for department budgets. Instead of manually building the table, they tell their agent: 'Create a new table called Q4 Budget with columns for Department Name (text) and Allocated Amount (number).' The agent uses create_table, defining the structure instantly.
Correcting Bad Data Entry
A team member realizes they accidentally typed 'Pending' instead of 'Complete' for a key milestone. They don't have to find the row, open it, and edit it. They simply ask: 'Update the status of Milestone ID 45 to Complete.' The agent calls update_row with precision.
Getting an Inventory Snapshot
The Ops team needs a count of all available products and their current stock levels. They ask: 'List all rows from the Inventory table, showing product name and stock.' The agent runs list_rows or query_sql, providing the immediate snapshot they need to report on.
The honest tradeoffs
Relying only on simple list functions
Asking 'Show me all tasks' using just list_rows returns every single field for every record, including massive text blocks and unnecessary metadata. This is huge noise.
Use the query_sql tool instead. Write a specific SELECT statement like SELECT task_name, status FROM Tasks WHERE status='open' to pull only the exact columns you need.
Manually checking column types
Assuming a field is an integer when it's actually a text string ('N/A'), leading to SQL syntax errors or incorrect query results.
Always start by running list_columns on the target table. This verifies every column name and its defined type before you write your complex query_sql.
Attempting multi-step updates in chat
Asking 'Change status to done, then update due date to tomorrow' as one single command. The agent might fail or only execute the first part.
Break it down into two specific commands. First: update_row to change the status. Second: a separate update_row call for the new due date. This ensures transactional integrity.
When It Fits, When It Doesn't
Use this server if your primary need is reliable, programmatic data interaction with structured records and tables. You're working in an environment where you frequently move from 'What does the data look like?' to 'Give me a calculated answer based on that data.' The strength of this tool set is its ability to handle both schema introspection (using list_tables, list_columns) and complex data manipulation (via query_sql).
Don't use it if you only need simple file storage or document management. If your goal is just generating content based on a prompt, a general-purpose LLM will suffice. You also shouldn't use it if you don't know the underlying data structure; always check get_base_metadata first. When in doubt about filtering, always default to writing an explicit query using query_sql rather than relying on simple list tools like list_rows, which are too broad.
Questions you might have
How do I check if a table exists before trying to query it using `query_sql`? +
You should first call list_tables. This returns an overview of all available databases and tables, confirming the name you need for your SQL statement.
Can I create a new field without manually updating the database schema? (using `create_row` or `update_row`) +
No. You must first use list_columns to see available fields, and if you need a truly new field, use create_table to define it.
What is the difference between `list_rows` and `query_sql`? +
list_rows gets every single row in a table—raw data dump. Use query_sql when you need specific filtering, grouping (GROUP BY), or aggregation (COUNT/AVG) that requires proper SQL syntax.
How do I make sure my AI client remembers the column names for `update_row`? +
Run list_columns first. The output gives you a list of every valid column name, which is what your agent needs to write an accurate update command.
When using `get_base_metadata`, how does it handle API token expiration or permission issues? +
It provides an explicit error code indicating authentication failure. The agent detects these specific errors and prompts you to check your SeaTable credentials or refresh the required access tokens.
What parameters do I need when calling `create_table` to ensure the new structure is usable? +
You must specify a unique name for the table. It's best practice to also define primary key columns and initial column types (text, number, etc.) during creation.
If I use `list_tables` first, does it limit my subsequent queries using `query_sql`? +
No, listing tables just provides a list of available databases. You still need to specify the target table name in your SQL query for it to run successfully.
Before running `delete_row`, is there a way to confirm that I'm actually deleting the correct record? +
The agent requires you to provide unique identifying details, like a row ID or combination of primary keys. This forces confirmation before the deletion command executes.
Can my AI automatically run SQL queries on my SeaTable data? +
Yes! Use the query_sql tool. Provide a standard SQL string (e.g., 'SELECT * FROM Tasks WHERE Status = "Done"'), and your agent will return the aggregate or filtered results instantly.
How do I find my SeaTable API Token? +
Open your SeaTable base, click the three dots (Advanced) next to the base name, select API Token, and create a permanent token for the base.
Does this integration work with self-hosted SeaTable instances? +
Yes! You can provide your custom serverUrl (e.g., https://seatable.mycompany.com) during setup to connect the MCP server to your private instance.
We've already built the connector for SeaTable. Just plug in your AI agents and start using Vinkius.
No hosting. No infrastructure. No complex setup.
All 11 tools are live and waiting.
You're up and running in seconds.
Gemini Vinkius gives your AI agents access to the full catalog of app connectors, all fully managed, secure, and enterprise-ready. One subscription, every tool you need.
Built, hosted, and secured by Vinkius. You just connect and go.