demandex-mcp
MCP server for Demandex e-commerce demand intelligence, providing Reddit-based opportunity cards and demand verdicts with free and paid per-call tools settled via x402/USDC.
README
demandex-mcp
MCP server for Demandex — pay-per-call e-commerce demand intelligence for AI agents. Demandex mines Reddit complaint/intent posts across ~70 communities into scored opportunity cards with verbatim evidence and permalinks, and answers ad-hoc demand verdicts for any physical-product query. Pay per call with USDC on Base via x402 — no signup, no accounts, no API keys.
- 2 free tools — no wallet, no API key.
- 5 paid tools — flat per-call USD prices, settled in USDC on Base Mainnet via x402.
Try before you pay: the free get_sample_opportunity tool returns a real, fixed opportunity
card in the same full shape as the paid get_opportunity, so you can learn the exact response
shape first.
Quick start
Free tier (no wallet)
{
"mcpServers": {
"demandex": {
"command": "npx",
"args": ["-y", "demandex-mcp"]
}
}
}
You get get_categories and get_sample_opportunity.
Paid tier (with wallet)
Add an EVM wallet private key (0x-prefixed) that holds USDC on Base Mainnet:
{
"mcpServers": {
"demandex": {
"command": "npx",
"args": ["-y", "demandex-mcp"],
"env": { "EVM_PRIVATE_KEY": "0x..." }
}
}
}
The paid tools then settle each call automatically.
Install snippets
Claude Code:
claude mcp add demandex -- npx -y demandex-mcp
# with a wallet:
claude mcp add demandex --env EVM_PRIVATE_KEY=0x... -- npx -y demandex-mcp
Claude Desktop / Cursor: add the JSON block above to your MCP config
(claude_desktop_config.json or .cursor/mcp.json).
Tools
Free tools
| Tool | Endpoint | Description |
|---|---|---|
get_categories |
GET /v1/categories |
Category index with opportunityCount, topScore, lastUpdated per category. |
get_sample_opportunity |
GET /v1/sample/opportunity |
Real, fixed sample opportunity card showing the full paid card shape. |
Paid tools
| Tool | Endpoint | Price (USD) | Description |
|---|---|---|---|
get_trending_opportunities |
GET /v1/opportunities/trending |
$0.01 | Top 20 trending cards as teasers (id, slug, title, category, score, signalCount, lastSeen). |
get_opportunities |
GET /v1/opportunities |
$0.02 | Up to 50 cards in a category (teaser + summary). Optional minScore (0-100). |
get_opportunity |
GET /v1/opportunity |
$0.05 | One FULL card by id or slug (exactly one) — all score components, productAngle, summary, verbatim evidence[]. |
gauge_demand |
POST /v1/gauge |
$0.03 | Demand verdict for a product query (3–200 chars), cached 7 days, corpus-only. Optional category. |
gauge_demand_live |
POST /v1/gauge/live |
$0.10 | Demand verdict with a LIVE Reddit search at request time (~45s budget). Same body as gauge_demand. |
Prices are flat per call — no bundles, no entitlements. The verdict shape (both gauges):
{ query, verdictScore (0-100), demandTemperature (cold|warm|hot), rationale, topEvidence[], relatedOpportunities[], cache, fetchedAt }.
How payments work (x402)
Demandex speaks the x402 exact scheme. The API's 402 advertises a
Base (USDC, EIP-3009) rail; this MCP settles on Base:
- The MCP requests a paid endpoint. With no payment, the API replies HTTP 402 with the accepted terms.
- If
EVM_PRIVATE_KEYis set, the MCP settles the call via the standard@x402/fetchV2 client and retries. If no key is set, the MCP returns the 402 price terms so you can see the cost without paying. - The API returns the data plus a
charged: truemarker.
Compute-first, settle-after: Demandex computes the full answer before charging, so you
are never billed for an error or an empty result (a still-warming index returns a no-charge
503 corpus_warming).
No accounts, no API keys — just a wallet with USDC on Base Mainnet. Physical products only.
Environment variables
| Variable | Required | Description |
|---|---|---|
EVM_PRIVATE_KEY |
No | 0x-prefixed wallet key with USDC on Base Mainnet. Required to pay the 5 paid tools on a live API. PRIVATE_KEY is accepted as a fallback. Free tools work without it; without a key the paid tools surface the 402 price terms. |
DEMANDEX_API_URL |
No | Override the API base URL (default https://api.demandex.dev). |
Disclaimer
Demandex provides informational e-commerce demand intelligence aggregated from public Reddit posts; physical products only; not professional or business advice; provided as-is. See https://api.demandex.dev/terms.txt.
Docs for agents: https://api.demandex.dev/llms.txt · full: https://api.demandex.dev/llms-full.txt
License
MIT © jcislo
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
Neon Database
MCP server for interacting with Neon Management API and databases
E2B
Using MCP to run code via e2b.
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.