pgEdge Launches Starfleet to Help AI Prototypes Reach Production

Developers frequently gravitate toward agile AI platforms for their ease of use, yet these systems often fail to meet the compliance, high-availability, and data sovereignty standards required for core business systems. pgEdge Starfleet offers a solution by providing a 100% Postgres-based environment that scales from a single instance to multi-region clusters, allowing teams to maintain identical architecture from initial experimentation through to full-scale deployment.

The platform integrates the pgEdge Agentic AI Toolkit, featuring an MCP server for connecting code generators like Claude Code and Cursor, alongside a RAG API server for retrieval-augmented generation. By utilizing true copy-on-write database branching, developers can run parallel experiments without the need for proprietary storage layers. Phillip Merrick, CEO of pgEdge, noted that the platform prevents developers from hitting architectural dead ends, offering the flexibility to deploy on pgEdge Cloud, private clouds, or air-gapped on-premises infrastructure.

Security remains a focal point of the release, with the platform incorporating IP allowlisting and SafeSession governance tools. Because the system is built on standard, open-source Postgres, it avoids vendor lock-in while providing the specific binaries required for hardware-validated enterprise environments. The service is available immediately, with a flat pricing model starting at $25 per month.