Coral

Coral is a data engine that lets AI agents query multiple data sources through a single SQL interface. Instead of making separate API calls to each service, agents can write SQL queries with JOINs across GitHub, Slack, Datadog, PagerDuty, Linear, and other sources. The tool handles authentication, pagination, rate limiting, and schema mapping automatically.

Designed for enterprise AI workflows, Coral reduces token usage and improves accuracy by letting agents retrieve exactly the data they need in one query rather than stitching together multiple tool calls. It runs locally as a desktop app or CLI, exposes itself as an MCP server for Claude and other AI frameworks, and learns relationships and recommended joins from usage patterns.