discovery — Claude Code MCP server: what it does, how to install
Connects AI agents to an open product catalog for contextual commerce. Lets assistants locate offers and place orders without proprietary ad rails.
claude mcp add --transport http discovery https://mcp.nexbid.dev/mcpWhat discovery does
The server connects coding assistants to the Nexbid commerce network through the open AdCP standard. Models search stores directly. Ranked results use a clear formula: 0.4 bid, 0.4 relevance, and 0.2 quality.
Core logic relies on Lean 4 formal verification with 47 verified theorems covering budgets and auctions. The system operates from Switzerland under digital opua GmbH without browser cookies.
All calls go to the remote endpoint https://mcp.nexbid.dev/mcp, so lost network access stops all searches immediately.
What it helps with
- locate commercial products by context and target geography
- evaluate marketplace items using a public scoring formula
- trigger purchases directly from an agent terminal
Example prompts for the agent
- Find organic olive oil options available for shipping to CH.
- Search advertiser products matching mechanical keyboards via discovery.
How to install discovery
Platform CLI: remote server
claude mcp add --transport http discovery https://mcp.nexbid.dev/mcpPlatform CLI: remote server
codex mcp add discovery --url https://mcp.nexbid.dev/mcpConfig file
{
"mcpServers": {
"discovery": {
"url": "https://mcp.nexbid.dev/mcp"
}
}
}Config file
{
"servers": {
"discovery": {
"type": "http",
"url": "https://mcp.nexbid.dev/mcp"
}
}
}Platform CLI: remote server
gemini mcp add -t http discovery https://mcp.nexbid.dev/mcpStep by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about discovery
How to install discovery in Claude Code?
Run claude mcp add --transport http discovery https://mcp.nexbid.dev/mcp in your terminal.
What does the discovery server do?
It grants AI models access to browse marketplace inventories and handle purchasing workflows over open protocols.
How does discovery score search results?
Products are ranked by multiplying 0.4 by the bid, 0.4 by relevance, and 0.2 by quality.