MCP integration
MCPPlug Amazon review data into MCP-ready agents such as WorkBuddy and Codex, with the same billing as our REST API.
One dataset, two ways in: use the REST API the traditional way (GET https://api.xdata-fly.com/v1/...), or plug the MCP endpoint below into your agents. Both share the same billing and rate-limit pipeline — a tool call over MCP costs exactly what an equivalent API request costs.
Endpoint & auth
The MCP endpoint speaks the standard Streamable HTTP transport and uses the same auth as our REST API: pass your key in the secret-key header.
https://api.xdatafly.com/mcpsecret-key: sk_live_xxxxxxxxxxxxxxxxAvailable tools
| Tool | Description | Status |
|---|---|---|
| amazon_reviews | Fetch Amazon reviews by ASIN: rating, title, body, masked author, verified purchase, helpful votes, date and country.params: asin (required), page_count, sort, star, site, include_media | Available |
| google_search | Structured Google SERP data.params: q, gl, hl, num, device | Coming soon |
Generic MCP client
For clients that support remote MCP (Cursor, VS Code, Cherry Studio and friends), add the following to the MCP config file.
{
"mcpServers": {
"xdata-fly": {
"url": "https://api.xdatafly.com/mcp",
"headers": {
"secret-key": "sk_live_xxxxxxxxxxxxxxxx"
}
}
}
}Connect Codex
Codex CLI reads MCP servers from ~/.codex/config.toml. Add the block below and restart Codex:
[mcp_servers.xdata-fly]
url = "https://api.xdatafly.com/mcp"
http_headers = { "secret-key" = "sk_live_xxxxxxxxxxxxxxxx" }Or add it with a single command:
codex mcp add xdata-fly --url https://api.xdatafly.com/mcp --header "secret-key=sk_live_xxxxxxxxxxxxxxxx"Connect WorkBuddy
WorkBuddy reads connectors from ~/.workbuddy/mcp.json (on Windows: C:\Users\<you>\.workbuddy\mcp.json). Add the server to mcpServers, save, and WorkBuddy can call the amazon_reviews tool in chat.
{
"mcpServers": {
"xdata-fly": {
"url": "https://api.xdatafly.com/mcp",
"headers": {
"secret-key": "sk_live_xxxxxxxxxxxxxxxx"
}
}
}
}If your WorkBuddy build only supports stdio connectors, use the next section instead.
stdio clients (bridge)
For stdio-only clients, bridge the remote endpoint into a local command with mcp-remote (Node.js required):
{
"mcpServers": {
"xdata-fly": {
"command": "npx",
"args": [
"-y",
"mcp-remote",
"https://api.xdatafly.com/mcp",
"--header",
"secret-key:sk_live_xxxxxxxxxxxxxxxx"
]
}
}
}Billing
MCP tool calls follow the same credit rules as the REST API: only successful calls are charged, and any failure (bad parameters, insufficient balance, upstream or parse error) releases the reserved credits automatically. Each response reports the credits charged and your remaining balance.
Ready to integrate?
Sign up for free trial credits — works with both REST and MCP.