Starter kits · Discovery first
MCP stacks that start with Discovery
Paste one of these into .cursor/mcp.json. Every kit includes
Influzer MCP Discovery so your agent can search the directory by capability —
then shortlist real servers before you commit secrets.
DISCOVERY ENDPOINT
https://www.influzer.ai/mcp/discovery
Developers in Cursor / Claude Code
Coding agent stack
Discover tools on demand, then keep GitHub + docs lookup in the repo. Discovery stays read-only; secrets stay in env.
“Search Influzer for MCP servers for a coding agent with GitHub and docs lookup”
.cursor/mcp.json
{
"mcpServers": {
"influzer-discovery": {
"url": "https://www.influzer.ai/mcp/discovery"
},
"github": {
"command": "npx",
"args": [
"-y",
"@modelcontextprotocol/server-github"
],
"env": {
"GITHUB_PERSONAL_ACCESS_TOKEN": "${env:GITHUB_TOKEN}"
}
},
"context7": {
"command": "npx",
"args": [
"-y",
"@upstash/context7-mcp"
]
}
}
}
Builders automating the web
Browser & scrape stack
Find scrape/crawl MCPs via Discovery, then pin a primary browser or scrape server once you shortlist.
“Search Influzer for MCP servers that can scrape URLs to markdown”
.cursor/mcp.json
{
"mcpServers": {
"influzer-discovery": {
"url": "https://www.influzer.ai/mcp/discovery"
},
"playwright": {
"command": "npx",
"args": [
"-y",
"@playwright/mcp@latest"
]
}
}
}
PMs and researchers in Claude / ChatGPT
Research & PM stack
Start with Discovery in chat connectors (no local install). Ask for topic guides and compare servers before approving writes.
“List Influzer topic guides for browser automation and recommend a shortlist”
.cursor/mcp.json
{
"mcpServers": {
"influzer-discovery": {
"url": "https://www.influzer.ai/mcp/discovery"
}
}
}
For Claude Desktop / ChatGPT, use the Setup guide tabs instead of this JSON — same Discovery URL.
Full setup (Claude · ChatGPT · Cursor · Claude Code) · Browse Top 100 · First MCP servers in Cursor