We build with Anthropic's Claude in two directions: inside our applications, where it helps people research and draft, and as a client of our data, through Model Context Protocol (MCP) connectors.
How we use it
- AI assistants that work with an application's own records, not just general knowledge.
- MCP connectors that let Claude call real services, so answers come from live data.
- Careful handling of data and prompts, with results people can check.
- AI-assisted development, which shortens delivery time on every project.
Where you will see it
- Summit, which includes a Claude-based research and drafting assistant.
- CDXGeoStream, whose MCP connector answers location questions inside Claude.
Planning a system that needs this stack?Talk to HFS