Hi! 👋 We are doing a big documentation refresh. Help us improve — what's missing or could be better? Let us know! Simply send an email or start a conversation in Google Groups!

Upcoming Tools

⚠Caution

The tools described on this page are planned for future ProxySQL v4.0 releases and are not yet available. Designs and interfaces are subject to change without notice.

Tools for the following endpoints are currently under development and will be available in future releases of ProxySQL v4.0.

Configuration Tools (/mcp/config)

Under Development — AI-driven management of ProxySQL global variables, servers, and users.

These tools will allow an AI agent or MCP client to read and modify ProxySQL’s runtime configuration without direct SQL access to the Admin interface. Planned capabilities include:

  • Reading and writing global variables (mysql-*, admin-*, pgsql-*)
  • Managing backend server entries (mysql_servers, pgsql_servers)
  • Managing user credentials (mysql_users, pgsql_users)
  • Managing query rules (mysql_query_rules)
  • Saving changes to disk and propagating them to a ProxySQL cluster

Admin Tools (/mcp/admin)

Under Development — Administrative tasks, cluster management, and system health checks.

These tools will expose ProxySQL’s operational administration capabilities to MCP clients. Planned capabilities include:

  • Triggering cluster re-synchronisation and checking cluster membership
  • Reloading certificates and other runtime assets
  • Graceful shutdown and restart signalling
  • Retrieving ProxySQL version and build information

Cache Tools (/mcp/cache)

Under Development — Inspection and invalidation of the Query Cache via natural language.

These tools will let AI agents interact with ProxySQL’s query result cache. Planned capabilities include:

  • Listing cached query digests and their hit counts
  • Invalidating specific cache entries by digest or pattern
  • Reporting cache memory usage and eviction statistics

Observation Tools (/mcp/observe)

Under Development — Real-time metrics and monitoring data analysis.

These tools will surface live operational data from ProxySQL’s stats tables to MCP clients, enabling AI-driven performance analysis. Planned capabilities include:

  • Retrieving current connection pool status per hostgroup
  • Streaming query latency histograms
  • Summarising slow-query logs and top query digests
  • Reporting backend health check outcomes

AI Tools (/mcp/ai)

Under Development — Direct bridge to LLMs and anomaly detection utilities.

These tools will provide higher-level AI primitives layered on top of ProxySQL’s observability data. Planned capabilities include:

  • Sending arbitrary prompts to a configured LLM backend with ProxySQL context pre-loaded
  • Detecting anomalous query patterns or connection behaviour
  • Generating natural-language explanations of ProxySQL configuration