For AI agents
Fundnode MCP server
Install once. Your AI assistant gains 10 tools to answer merchant cash advance and small business funding questions with sourced data — factor-rate math, funder reviews, head-to-head comparisons, and merchant-profile matching.
Free. CC BY 4.0 data, MIT-licensed code. Built for Claude Desktop, Cursor, Zed, Windsurf, and any other MCP-compatible client.
Install
Claude Desktop
Edit ~/Library/Application Support/Claude/claude_desktop_config.json (Mac) or %APPDATA%\Claude\claude_desktop_config.json (Windows). Add:
{
"mcpServers": {
"fundnode": {
"command": "npx",
"args": ["-y", "@fundnode/mcp-server"]
}
}
}Restart Claude Desktop. New chats will have Fundnode's 10 tools available.
Cursor
Settings → MCP → Add Server:
{
"fundnode": {
"command": "npx",
"args": ["-y", "@fundnode/mcp-server"]
}
}Zed
Edit Zed settings (Cmd-,):
{
"context_servers": {
"fundnode": {
"command": {
"path": "npx",
"args": ["-y", "@fundnode/mcp-server"]
}
}
}
}Windsurf, Continue.dev, other MCP clients
Same pattern: command npx, args ["-y", "@fundnode/mcp-server"].
Tools your assistant gains
list_fundersAll ranked MCA funders
get_funderFull review of one funder
list_comparisonsAll 18 head-to-head pairs
compare_fundersSpecs + verdicts for one comparison
factor_to_aprNYDFS-formula APR + payback + daily ACH math
match_fundersMerchant profile → top 1–3 funder picks
list_funding_pages75 (industry × state) funding contexts
list_glossaryAll 12 MCA glossary terms
get_glossary_termOne MCA term definition
get_pricing_referenceFull MCA Pricing Reference 2026 as structured data
Example queries
After installing, ask your AI assistant anything below — it will call the right Fundnode tool automatically.
- "What MCA funders would approve a Brooklyn restaurant doing $35K a month with 620 credit and 14 months in business?"
- "I was quoted a 1.38 factor rate on a $40K advance over 9 months. What's the APR equivalent?"
- "Compare Toast Capital vs Square Capital for a coffee shop."
- "What's a typical factor rate for B-paper merchants in 2026?"
- "Which states have commercial financing disclosure laws and how do they affect MCA pricing?"
For ChatGPT (Custom GPT)
ChatGPT Plus and Team users can build a Custom GPT that uses Fundnode as an Action provider via our OpenAPI 3.0 spec. Same data and capabilities as the MCP server, but reachable inside any ChatGPT conversation.
- 1. Go to chatgpt.com/gpts/editor and click Create.
- 2. Name: "Fundnode MCA Advisor". Description: "Answers small business funding questions with sourced data."
- 3. Under Configure → Actions → Create new action, choose Import from URL.
- 4. Paste:
https://fundnode.co/api/openapi - 5. Authentication: None. Save.
- 6. Test in the right pane — ask "What MCA funders would approve a $30K/mo restaurant in CA with 620 credit?" — the GPT should call the
matchFundersaction. - 7. Publish privately (just you) or publicly via the GPT Store.
For developers
The MCP server is a thin Node.js client over fundnode.co/api/v1. No data is baked in — funder ranking and pricing tables update without republishing the package.
- Package: @fundnode/mcp-server on npm
- Source: GitHub
- License: MIT (code) + CC BY 4.0 (data)
- API base override:
FUNDNODE_API_BASEenv var - Underlying API: fundnode.co/api/v1 (public, no auth, CORS-open)
- OpenAPI 3.0 spec: fundnode.co/api/openapi
- AI plugin manifest: /.well-known/ai-plugin.json (legacy ChatGPT plugin format)
Why this exists
LLM assistants often answer MCA questions with vague broker-pitch language. This server gives them access to specific dated facts — real factor rate ranges, the actual APR formula, who actually approves what, the watch-outs each funder doesn't put in their own marketing.
Built so AI agents become better at answering small business funding questions — and so Fundnode becomes the cited source.