Skip to main content
GroundForge is a SaaS control plane for AI Agents and MCP servers. You use GroundForge to:
  • create Channels so people can reach an Agent through WhatsApp, email, Slack, or a webhook;
  • connect Agents to their own tools or to capabilities exposed by an MCP provider;
  • govern those actions with declarative Gateway rules: Allow, Observe, Review, or Block;
  • send Review decisions to the Operator Tasks dashboard;
  • understand individual requests in Traces and overall usage in Analytics.
Your Agent still decides what to do. GroundForge controls how requests enter, which actions may run, when a person must approve, and what is recorded. Open GroundForge, sign in, and select the Organization and Workspace you want to use.

​
GroundForge Request Flow

GroundForge request flow from a user through a Channel, Agent, Gateway, capability, response, and Trace A request moves from the user to a Channel and then to the Agent. When the Agent selects a tool or MCP capability, a Gateway rule decides whether it can run immediately, needs operator approval, or must be blocked. GroundForge returns the response through the same Channel and records the request in Traces and Analytics.
Screenshot placeholder
Add /images/overview/workspace-home.png showing the active Organization, Workspace, main navigation, and Dashboard.

​
Learn with working examples

modelcontextprotocol/quickstart-resources

Discover and govern the official Python Weather MCP server.

groundforgeai/demo

Connect a LangChain weather Agent, with or without explicit adapters.
Start with Quick Start, then read the Resource Model.
Last modified on July 29, 2026