Caspio
Server Details
Your Caspio data is MCP-enabled. Bring any dataset and put your AI assistants to work. No coding.
- Status
- Healthy
- Uptime
- 100.0% over the last 24 hours
- OAuth
- Works in Glama
- Last Tested
- Transport
- Streamable HTTP · MCP 2025-11-25
- URL
- ChatGPT
- –
- Claude
- –
TDQS
Scored across 15 tools
Each tool targets a distinct resource type and action: table vs view operations are clearly separated, discovery tools cover tables/views/directories, and file metadata tools are differentiated by attachment vs. file system. Where overlap could exist (table vs. view), the descriptions explicitly define the difference.
All tools follow a consistent verb_noun snake_case pattern (discover_tables, query_view, insert_table_record, delete_view_records, get_file_metadata). The table/view variants mirror each other exactly, making the set highly predictable.
15 tools is well-scoped for a database platform covering tables, views, directories, and files. Each tool serves a distinct purpose and earns its place without redundancy or bloat.
Core CRUD operations are complete for both tables and views, with discovery support for all schema types. The main gap is file handling: search and metadata retrieval exist, but there is no download/upload tool, despite download_file being referenced in search_files.
Publisher details
- Operator
- Caspio
- Operator website
- https://www.caspio.com/
- Vendor relationship
- First-party · Publisher source
- Privacy policy
- Not provided
- Trust center
- https://trust.caspio.com/
- Restrictions
- Not available
Available Tools
15 toolsdelete_table_recordsADestructiveInspect
Deletes one or more records from a Caspio table that match the WHERE clause. This is irreversible. WHERE is required and supports Microsoft SQL Server syntax. A broad WHERE clause will delete all matching records. To delete a single record, filter by PK_ID or another unique field. Returns the count of deleted records. IMPORTANT: field names in WHERE are case-insensitive and must come from discover_tables' field list -- never guess. Call discover_tables first for any table not already discovered earlier in this conversation to get its tableId.
Inputs: where (required), tableId (required).
delete_view_recordsADestructiveInspect
Deletes one or more records through a Caspio view that match the WHERE clause. This is irreversible. In multi-table views, only records from the primary table are deleted. WHERE is required and supports Microsoft SQL Server syntax. A broad WHERE clause will delete all matching records. To delete a single record, filter by PK_ID or another unique field. Returns the count of deleted records. IMPORTANT: field names in WHERE are case-insensitive and must come from discover_views' field list -- never guess. Call discover_views first for any view not already discovered earlier in this conversation to get its viewId.
Inputs: where (required), viewId (required).
discover_directoriesARead-onlyInspect
Discovers directories in the Caspio account along with their field schemas in a single call. Directories are identity and authentication services that manage app user accounts, supporting Caspio-native authentication as well as SAML-based federation (inbound from providers like Okta or Active Directory, and outbound to external systems). A Caspio account can have multiple directories for different user types (e.g., one for employees, another for customers). Returns each matching directory's id (needed for query_directory_users), name, and field definitions. Use nameFilter to search for a specific directory, or omit it to list all directories.
Inputs: pageSize, nameFilter, pageNumber, includeTriggeredActions.
discover_tablesARead-onlyInspect
Discovers tables in the Caspio account along with their field schemas in a single call. Tables are the primary data storage objects in Caspio (similar to database tables); table names are customer-defined and reflect their specific use case (e.g., Patients, ParkingPermits, Inventory_Items). Returns each matching table's id (required by query_table, insert_table_record, update_table_records, delete_table_records, and get_attachment_metadata), name, and field definitions. Each field includes: name (the identifier used in all queries), dataType, label (optional friendly display name), unique, editable (false for auto-generated fields like Autonumber, Timestamp, GUID, Random ID, and Formula -- do not include these in insert or update payloads), and description. Caspio API data types: STRING (short text, max 255 chars), TEXT (long text, max 64K chars), NUMBER (decimal), INTEGER (whole number), CURRENCY, DATE/TIME, YES/NO (boolean -- use 1 for true and 0 for false in queries), FILE (legacy, referenced separate file object), ATTACHMENT, LIST-STRING / LIST-NUMBER / LIST-DATE/TIME, PASSWORD, TIMESTAMP, RANDOM ID, AUTONUMBER, PREFIXED AUTONUMBER, GUID. Use field name (never label) in SELECT, WHERE, ORDER BY, and GROUP BY clauses. Use nameFilter to search for a specific table, or omit it to list all tables.
Inputs: pageSize, nameFilter, pageNumber, includeTriggeredActions.
discover_viewsARead-onlyInspect
Discovers views in the Caspio account along with their field schemas in a single call. Views are predefined data sources that can filter a single table or join multiple tables; view names are customer-defined. Tables are the primary data objects in Caspio; use views when the user's request aligns with an existing view or when the needed data spans multiple tables. Returns each matching view's id (required by query_view, insert_view_record, update_view_records, and delete_view_records), name, and field definitions. Each field includes: name (the identifier to use in queries), tableFieldName (the source table and field in Table.Field format, revealing which underlying table each field comes from), dataType, label, unique, editable (false for auto-generated fields and for fields from non-primary joined tables in multi-table views), and description. Data types are the same as table fields: STRING, TEXT, NUMBER, INTEGER, CURRENCY, DATE/TIME, YES/NO (boolean -- use 1 for true and 0 for false in queries), FILE (legacy), ATTACHMENT, LIST-STRING / LIST-NUMBER / LIST-DATE/TIME, PASSWORD, TIMESTAMP, RANDOM ID, AUTONUMBER, PREFIXED AUTONUMBER, GUID. Use field name (never label or tableFieldName) in SELECT, WHERE, ORDER BY, and GROUP BY clauses. Use nameFilter to search for a specific view, or omit it to list all views.
Inputs: pageSize, nameFilter, pageNumber, includeTriggeredActions.
get_attachment_metadataBRead-onlyInspect
Returns metadata for files stored in an Attachment-type field of a Caspio table. Returns PK_ID (the table record's internal ID), FileName, FileType, Size, DateCreated, and DateUpdated. Use the optional WHERE clause to filter to specific records. A table record can have multiple Attachment-type fields.
Inputs: tableId (required), attachmentFieldName (required), where.
get_file_metadataARead-onlyInspect
Returns metadata for a file in Caspio's file system, including its ExternalKey (an internal GUID). Also returns Name, Size, ContentType, DateCreated, and LastModified. The file path comes from a File-type field value in a table or view record (e.g., '/photos/logo.jpg').
Inputs: filePath (required).
insert_table_recordAInspect
Inserts a single record into a Caspio table. The payload is a flat JSON object of field name/value pairs (e.g., {'Name': 'John', 'Age': 30}). Only include editable fields -- do not include auto-generated fields (Autonumber, Timestamp, Random ID, GUID, Prefixed Autonumber, Formula, or PK_ID). Returns the full created record including PK_ID (a system-generated unique identifier present on every Caspio table record). IMPORTANT: field names are case-insensitive and must come from discover_tables' field list -- never guess. Call discover_tables first for any table not already discovered earlier in this conversation to get its tableId and confirm which fields exist and which are editable.
Inputs: payload (required), tableId (required).
insert_view_recordAInspect
Inserts a single record through a Caspio view. The payload is a flat JSON object of field name/value pairs (e.g., {'SaleStatus': 'Open', 'SaleAmount': 500}). Only include editable fields -- in multi-table views, only fields from the primary table are editable (check Editable=true in the view's field definitions). Do not include auto-generated fields (Autonumber, Timestamp, Random ID, GUID, Prefixed Autonumber, Formula, or PK_ID). Returns the full created record including PK_ID. IMPORTANT: field names are case-insensitive and must come from discover_views' field list -- never guess. Call discover_views first for any view not already discovered earlier in this conversation to get its viewId and confirm which fields exist and which are editable.
Inputs: viewId (required), payload (required).
query_directory_usersARead-onlyInspect
Queries users from a Caspio directory. Directories are tables with authentication and identity capabilities enabled, supporting Caspio-native authentication and SAML-based federation. This tool returns only the directory-designated fields plus three system attributes not available through regular table queries: _status ('Active', 'Inactive', 'ActivationPending', 'ActivationExpired', 'Locked', 'Suspended', 'PasswordExpired'), _sign_in_method (e.g., 'Caspio' or an external SAML provider), and _2fa_status ('Enabled', 'Disabled', 'Unavailable'). These system attributes can be used in SELECT, WHERE, ORDER BY, and GROUP BY clauses. The underlying table may contain additional fields (e.g., photo, date of birth, employment dates) that are only accessible through query_table using that table's Id (from discover_tables). IMPORTANT: field names are case-insensitive and must come from discover_directories' field list -- never guess. Call discover_directories first for any directory not already discovered earlier in this conversation to get its directoryId.
Inputs: select (required), directoryId (required), limit, where, +5 more.
query_tableARead-onlyInspect
Queries records from a Caspio table. The backend supports Microsoft SQL Server WHERE syntax (LIKE, AND, OR, NOT, IN, BETWEEN, comparison operators, etc.). Returns a JSON object with a 'records' array of flat key-value objects where keys are field names, and optional 'pagination' metadata when getPaginationInfo is true. Maximum 1000 records per request (default 100). For larger result sets, use pageNumber and pageSize to paginate. IMPORTANT: field names are case-insensitive and must come from discover_tables' field list -- never guess a field name from its label or from naming conventions. Call discover_tables first for any table not already discovered earlier in this conversation to get its tableId.
Inputs: select (required), tableId (required), limit, where, +5 more.
query_viewARead-onlyInspect
Queries records from a Caspio view. Views are predefined data sources that can filter a single table or join multiple tables. The query syntax is identical to querying tables -- the backend supports Microsoft SQL Server WHERE syntax (LIKE, AND, OR, NOT, IN, BETWEEN, comparison operators, etc.). Returns a JSON object with a 'records' array of flat key-value objects where keys are field names, and optional 'pagination' metadata when getPaginationInfo is true. Maximum 1000 records per request (default 100). For larger result sets, use pageNumber and pageSize to paginate. IMPORTANT: field names are case-insensitive and must come from discover_views' field list -- never guess a field name from its label or from naming conventions. Call discover_views first for any view not already discovered earlier in this conversation to get its viewId.
Inputs: select (required), viewId (required), limit, where, +5 more.
search_filesARead-onlyInspect
Searches for files in Caspio's file system by name. Returns matching files' metadata (path, name, size, content type, dates) usable with get_file_metadata or download_file. Use this to find a file when the exact path isn't already known (e.g., from a File-type field value).
Inputs: name, pageSize, sortField, pageNumber, +1 more.
update_table_recordsAInspect
Updates one or more records in a Caspio table that match the WHERE clause. The payload is a flat JSON object containing only the fields to update (e.g., {'Status': 'Closed', 'ClosedDate': '2026-01-15'}). Fields not included in the payload remain unchanged. WHERE is required and supports Microsoft SQL Server syntax. Be cautious: a broad WHERE clause will update all matching records. To update a single record, filter by PK_ID or another unique field. Returns all updated records. IMPORTANT: field names are case-insensitive and must come from discover_tables' field list -- never guess. Call discover_tables first for any table not already discovered earlier in this conversation to get its tableId.
Inputs: where (required), payload (required), tableId (required).
update_view_recordsAInspect
Updates one or more records through a Caspio view that match the WHERE clause. The payload is a flat JSON object containing only the fields to update (e.g., {'SaleStatus': 'Closed', 'SaleAmount': 1000}). Fields not included in the payload remain unchanged. In multi-table views, only fields from the primary table can be updated (check Editable=true in the view's field definitions). WHERE is required and supports Microsoft SQL Server syntax. A broad WHERE clause will update all matching records. To update a single record, filter by PK_ID or another unique field. Returns all updated records. IMPORTANT: field names are case-insensitive and must come from discover_views' field list -- never guess. Call discover_views first for any view not already discovered earlier in this conversation to get its viewId.
Inputs: where (required), viewId (required), payload (required).
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Related MCP Connectors
Build and run AI workflows, agents, data tables and 1000+ API integrations from any MCP client.
Cloud-hosted MCP server for secure AI access to enterprise data sources via CData Connect AI.
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
- AlicenseAqualityDmaintenanceGive your AI agent safe, plain-English access to any database via MCP. Ask questions in natural language, get SQL queries and results, run read-only queries, and set up scheduled alerts.934 npm1MIT
- AlicenseNot gradedqualityCmaintenanceAuto-generates MCP servers from any data source, enabling AI clients to query databases, spreadsheets, and APIs with zero code.29 PyPI5MIT

AgenticBIofficial
FlicenseNot gradedqualityDmaintenanceConnect your business data to any MCP client. AgenticBI agents find the right data, join across systems, run queries, build dashboards, generate reports, and automate analytics workflows. No SQL required.-- AlicenseNot gradedqualityDmaintenanceEnables AI assistants to query and analyze databases using natural language through MCP and HTTP API. Supports 17+ databases and integrates with 55+ platforms.131 npmMIT
Glama MCP Gateway
Add one secure layer between your agents and this server.