KnowMint MCP Server
AI agent knowledge marketplace MCP server. Agents autonomously discover, purchase (x402/Solana), and retrieve human experiential knowledge.
KnowMint MCP Server
The first knowledge marketplace where AI agents pay directly with SOL — autonomously via x402.
Connect Claude Code, Cursor, or OpenCode to KnowMint. Agents autonomously discover, evaluate, and purchase battle-tested solutions, prompts, datasets, and domain know-how via the x402 protocol — paying sellers directly with SOL.
Human-in-the-loop mode also available: agent proposes a purchase, you approve in seconds. Your choice of autonomy level.
Quickstart (npx — no install needed)
Claude Code (~/.claude/mcp.json)
{
"mcpServers": {
"knowmint": {
"command": "npx",
"args": ["--yes", "--package", "@knowmint/[email protected]", "mcp-server"],
"env": {
"KM_API_KEY": "km_xxx",
"KM_BASE_URL": "https://knowmint.shop"
}
}
}
}
Cursor (~/.cursor/mcp.json)
{
"mcpServers": {
"knowmint": {
"command": "npx",
"args": ["--yes", "--package", "@knowmint/[email protected]", "mcp-server"],
"env": {
"KM_API_KEY": "km_xxx",
"KM_BASE_URL": "https://knowmint.shop"
}
}
}
}
OpenCode (.opencode/config.json)
{
"mcp": {
"servers": {
"knowmint": {
"command": "npx",
"args": ["--yes", "--package", "@knowmint/[email protected]", "mcp-server"],
"env": { "KM_API_KEY": "km_xxx", "KM_BASE_URL": "https://knowmint.shop" }
}
}
}
}
Claude Desktop (~/Library/Application Support/Claude/claude_desktop_config.json)
{
"mcpServers": {
"knowmint": {
"command": "npx",
"args": ["--yes", "--package", "@knowmint/[email protected]", "mcp-server"],
"env": {
"KM_API_KEY": "km_xxx",
"KM_BASE_URL": "https://knowmint.shop"
}
}
}
}
Windows:
%APPDATA%\Claude\claude_desktop_config.jsonKM_BASE_URLdefaults tohttps://knowmint.shopwhen omitted. Get your API key at knowmint.shop/settings/api or runkm login.
Available Tools
| Tool | Description |
|---|---|
km_search | Search knowledge listings by keyword, type, or price |
km_get_detail | Get full metadata + preview for a listing |
km_purchase | Record a purchase after sending payment (tx_hash) |
km_get_content | Retrieve purchased full content |
km_publish | Publish a new knowledge listing |
Authentication
Priority order:
KM_API_KEYenvironment variable~/.km/config.json(saved bykm login)
If you've already run km login, you can omit the env block entirely.
Autonomous Purchase Flow (x402 — no external wallet MCP needed)
Agents with their own wallet (e.g. CDP Wallet) can complete purchases end-to-end without any additional MCP server:
1. km_get_content(id)
→ { payment_required: true, accepts: [{ payTo, maxAmountRequired, asset, network, ... }] }
2. Agent sends SOL/USDC from its own wallet
→ obtain tx_hash / signature
3. Build payment_proof:
base64url(JSON.stringify({
scheme: "exact",
network: "<value from accepts[].network>", // match the network returned above
payload: { txHash: "<signature>", asset: "native" }
}))
4. km_get_content(id, payment_proof: "<base64url>")
→ { full_content: "...", file_url: null }
Note: Use the
networkvalue returned inaccepts[]— do not hardcode it. Devnet example:"solana:EtWTRABZaYq6iMfeYKouRu166VU2xqa1"
Manual purchase via Phantom MCP also works: send tx → km_purchase(id, tx_hash) → km_get_content(id).
Local Development Setup
cd mcp && npm install && npm run build
{
"mcpServers": {
"knowmint": {
"command": "node",
"args": ["/path/to/knowledge_market/mcp/dist/index.js"],
"env": { "KM_API_KEY": "km_xxx" }
}
}
}
Dev mode (TypeScript, no build step)
{
"mcpServers": {
"knowmint": {
"command": "node",
"args": ["--experimental-strip-types", "/path/to/knowledge_market/mcp/src/index.ts"],
"env": { "KM_API_KEY": "km_xxx" }
}
}
}
Verify the server starts
# Should print error to stderr and exit (no KM_API_KEY set)
npx --yes --package @knowmint/[email protected] mcp-server
# With key — should stay running (waiting for MCP messages on stdin)
KM_API_KEY=km_xxx KM_BASE_URL=http://localhost:3000 npm run dev
# Interactive inspection
npx @modelcontextprotocol/inspector node --experimental-strip-types mcp/src/index.ts
Requirements
- Node.js ≥ 22.6.0
- KnowMint API key (get one here)
相关服务器
MCP Trakt
Access real-time entertainment data and personal viewing history from the Trakt.tv API.
WhoisFreaks
A Model Context Protocol (MCP) server that exposes the full WhoisFreaks API suite as AI-callable tools. Works with Claude Desktop, Cursor, Windsurf, VS Code, Continue, Zed, and any other MCP-compatible AI client.
Memory Anchor
Persistent memory MCP server for AI coding assistants with 5-layer cognitive model, hybrid semantic+keyword search, and checkpoint/resume. Zero cloud, local-first.
Hidden Empire
Play a legendary text adventure by talking to your AI — no commands to memorize. The Hidden Empire puts a full underground world of puzzles, treasures, and trolls inside your conversation. Speak naturally: say 'head north,' 'grab the lantern,' or 'what am I carrying?' and your AI handles the rest. Execute multi-move plans in one shot, undo mistakes instantly, and save up to 20 named playthroughs you can resume from any session. Based on the MIT-licensed Zork I source, rebuilt from the ground up for AI-native play.
Solematica
Solar estimates, provider comparison, energy prices and blog for the Italian market. 6 tools, 5 free. Compare providers, analyze roofs via Google Solar API, check PUN/ARERA prices.
AI Compliance Monitor
Structured regulatory intelligence for AI agents. Check compliance obligations, deadlines, and cross-jurisdiction requirements for EU AI Act, Singapore IMDA, and Colorado AI Act.
ReadyPermit
AI-powered property zoning, buildability, and ADU eligibility analysis for any U.S. address via 20+ government data sources.
Neonia
The ultimate platform for Autonomous AI Agents. Features include Autonomous Tool Discovery (dynamically finds and executes missing capabilities), Stateful Cloud Memory (remembers cross-session context), Context Packing (saves LLM tokens), and 20+ specialized dev tools.
openterms-mcp
Ed25519-signed consent receipts + programmable policy engine for AI agents before they take actions.
Speech AI
Production speech AI MCP server with pronunciation scoring, speech-to-text, and text-to-speech — 10 tools, 7 resources, 3 prompts.