Tiger CLI MCP Server – Tiger Cloud Database Tools
Tiger CLI includes a Model Context Protocol server that lets AI assistants interact with Tiger Cloud infrastructure and database services. It supports service lifecycle operations, database query execution, and optional proxied documentation search for PostgreSQL, TimescaleDB, and Tiger Cloud guidance.
Overview
Tiger CLI is the official command-line interface for Tiger Cloud, published by Tiger Data in the timescale/tiger-cli repository. In addition to normal CLI commands for authentication, service management, database connectivity, and configuration, Tiger CLI includes an integrated MCP server for AI assistants. The MCP server is intended for clients such as Claude Code, Codex, Cursor, Gemini CLI, VS Code, and Windsurf, and it enables an assistant to work with Tiger Cloud resources through documented tools instead of relying only on static model knowledge.
What the MCP server enables
The Tiger MCP server exposes tools for Tiger Cloud service management and database operations. Documented service tools include listing services, retrieving service details, creating services, forking services, starting or stopping services, resizing CPU and memory allocation, updating a service master password, and viewing service logs. It also exposes db_execute_query, which executes SQL queries against a database service with support for parameterized queries, custom timeouts, and connection pooling.
The server can also proxy tools from Tiger Data's documentation MCP server. When enabled, this adds view_skill and search_docs capabilities for PostgreSQL, TimescaleDB, and Tiger Cloud documentation, helping coding agents retrieve implementation guidance while working on database-related tasks.
When to use it
Use Tiger CLI MCP when an AI assistant should help operate Tiger Cloud services or inspect and query Tiger Cloud databases. Practical workflows include listing available services before a migration, creating or forking a service for development, checking logs during troubleshooting, generating safe SQL queries, and searching Tiger Cloud or PostgreSQL documentation while implementing database features.
Connection and authentication
Tiger CLI supports local stdio MCP usage through tiger mcp start, and the interactive installer configures supported MCP clients with tiger mcp install. It also supports Streamable HTTP by starting the server with tiger mcp start http --port 8080 and connecting an MCP client to http://localhost:8080. Authentication is handled through Tiger CLI's existing authentication and configuration, so users should sign in with tiger auth login before using the MCP server. For noninteractive environments, TIGER_PUBLIC_KEY and TIGER_SECRET_KEY can override stored credentials.
Key considerations
The MCP server can perform mutating service operations, so permission and safety controls matter. Tiger CLI documents a read_only configuration option, also available as TIGER_READ_ONLY, that refuses write or destructive CLI commands and their MCP equivalents. The documentation proxy is enabled by default and can be disabled with tiger config set docs_mcp false or equivalent configuration. Store credentials securely, limit access to trusted MCP clients, and use read-only mode when an assistant only needs inspection or query support.
Supported Transports
stdio
Command: tiger
Args:
mcpstart
streamable_http
URL: http://localhost:8080
Command: tiger
Args:
mcpstarthttp--port8080
Frequently Asked Questions
- When should an AI agent use the Tiger CLI MCP server?
- Use it when an assistant needs to manage Tiger Cloud database services, inspect service status, retrieve service logs, execute SQL against a selected service, or consult PostgreSQL, TimescaleDB, and Tiger Cloud documentation while helping with database development.
- What does this MCP server add to an AI agent's capabilities?
- It gives the agent live tool access to Tiger Cloud service management and database operations through Tiger CLI, including service lifecycle tools and the documented `db_execute_query` tool, rather than relying only on general database knowledge.
- What can the AI agent access or manage through this server?
- Documented MCP tools cover listing services, reading service details, creating and forking services, starting and stopping services, resizing resources, updating a service master password, viewing logs, and executing SQL queries. When the docs proxy is enabled, the agent can also use `view_skill` and `search_docs` for Tiger Data documentation.
- How is authentication configured for this MCP server?
- The server uses Tiger CLI authentication and configuration. Users normally run `tiger auth login` first, while environment variables such as `TIGER_PUBLIC_KEY` and `TIGER_SECRET_KEY` can take priority over stored credentials for noninteractive setups. Secrets should be supplied through environment variables and never embedded directly in client configuration.
- Which transport should be used for this MCP server?
- Use stdio with `tiger mcp start` for local MCP clients configured by `tiger mcp install` or manual client configuration. Use Streamable HTTP when a client needs an HTTP endpoint; Tiger CLI documents starting it with `tiger mcp start http --port 8080` and connecting to `http://localhost:8080`.