Database
Query your unified profile with PostgreSQL
Query your unified, identity-resolved customer profiles directly from PostgreSQL. The same data your AI agents read over MCP—available to your warehouse, BI tools, and SQL-native workflows.
Enable access. Paste a connection string. Run a query.
Create an account at pathbound.ai. Free during public alpha.
In the dashboard, go to Integrations → PostgreSQL → Connect. Pathbound provisions a read-only connection to your unified, identity-resolved customer data.
Paste the connection string into psql, Metabase, Looker, dbt, or any Postgres-compatible client. The unified contacts, companies, and events are available as tables.
psql "postgresql://USER:[email protected]:5432/your_workspace"Try SELECT email, lifecycle_stage, source_count FROM contacts WHERE lifecycle_stage = 'trial';—the unified profile, queryable, with sources already merged.
Not every consumer of customer data is an AI agent. Pathbound exposes the same unified, identity-resolved profiles over PostgreSQL—so your BI tools, dashboards, and SQL pipelines read exactly what your agents do, without building a second integration.
Agents query over MCP; analysts and pipelines query over SQL. Same identity resolution, same merged profiles, same source of truth—just a different door.
Contacts and companies arrive already deduped and stitched across every source. The joins and reconciliation you’d otherwise hand-build are done before the query runs.
A read-only view of the unified data. Querying can’t change your records—point any tool at it without risk.
Anything that speaks Postgres connects directly—Metabase, Looker, dbt, a notebook, or a SQL-native agent. No new SDK, no new format.
Yes. The PostgreSQL surface is a read-only view of your unified customer data—querying never changes your records or your connected source tools.
This is the egress side: it exposes Pathbound’s already-unified, identity-resolved profiles for you to query over SQL. It’s for consumers—BI, dashboards, pipelines—not for ingesting a new source.
Use MCP for AI agents that act conversationally on a customer; use SQL for analytics, dashboards, batch pipelines, and tools that already speak Postgres. Both read the same unified profiles.
The unified contacts, companies, and events—identity-resolved across every connected source, so you query one clean model instead of joining raw per-tool exports.
Tenant-isolated with scoped, read-only credentials and encryption at rest via our infrastructure provider’s default configuration. GDPR-aligned; SOC 2 readiness in progress. Pathbound does not sell customer data.
Same unified data—just a different surface. Connect your sources, then read them over MCP or SQL. See the integrations directory.
Sign up, connect one CRM or drop the events snippet, and point your MCP client at mcp.pathbound.ai/mcp.