Reach Dog MCP
OfficialReach Dog MCP server
Classify products, group catalogs, research markets and create content from an MCP client.
Install and connect
Use Node.js 24 or newer. Sign in to the Reach Dog developer console, create your account's API key, and use your existing tokens. The console documents the scopes enabled on your key.
Add this to the MCP configuration in Claude Desktop or Cursor, replacing the API key and account label with your own values.
{
"mcpServers": {
"reachdog": {
"command": "npx",
"args": ["--yes", "@reachdog/mcp@0.1.0"],
"env": {
"REACHDOG_API_URL": "https://api.reach.dog",
"REACHDOG_API_KEY": "YOUR_API_KEY",
"REACHDOG_ACCOUNT_LABEL": "my-application"
}
}
}
}For VS Code, put the reachdog object under servers in .vscode/mcp.json and add "type": "stdio" to that object.
The tools are get_usage, list_parameters, estimate_cost, enrich_title, research_market, generate_content, enrich_catalog, get_catalog_status, and retry_payment.
Related MCP server: Agentic Product Protocol MCP Server
First request
Ask your client to enrich one of your product titles with product_type. Review the estimate and approve the request through the client's approval dialog. Catalog enrichment first returns a quote, then requires acceptance of that exact quote. Each request uses your Reach Dog account's tokens.
The session budget counts Reach Dog billing tokens. Set REACHDOG_MAX_TOKENS_PER_SESSION to a whole number to limit estimated spending across the process. A client with elicitation prompts for each payment. In a client without elicitation, configuring a cap authorizes spending up to the cap when the model confirms. With neither approval mechanism only previews are available.
Build from source
Run npm ci, npm run build, and npm pack from this repository. To use a source build, configure the MCP command as node with the absolute path to dist/server.js as its argument. npm test runs the offline client and tool checks.
Support
Use GitHub issues for bugs. Include the version and error code, keeping API keys and private product records out of the report. See the API reference for request fields, scopes and errors.
MIT licensed. API usage uses your Reach Dog account and its available tokens.
Available Tools
9 toolsenrich_catalogC
Submit a catalog for a quote with body. Accept a previously displayed quote with batch_id, acceptance and confirm=true.
Inputs: body, confirm, batch_id, run_label, +1 more.
enrich_titleC
Preview by default. Set confirm to request approval for this exact body.
Inputs: body (required), confirm, run_label.
estimate_costC
Estimate one title, research or content request in Reach Dog tokens.
Inputs: parameters (required).
generate_contentC
Preview by default. Set confirm to request approval for this exact body.
Inputs: body (required), confirm, run_label.
get_catalog_statusC
Read current catalog status and settled results.
Inputs: batch_id (required).
get_usageB
Read account usage and current unit price.
Inputs: no named parameters.
list_parametersC
List available API parameters, grouped by endpoint.
Inputs: no named parameters.
research_marketC
Preview by default. Set confirm to request approval for this exact body.
Inputs: body (required), confirm, run_label.
retry_paymentB
Request approval to retry a refused payment for the original accepted quote.
Inputs: batch_id (required), confirm.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 9 tools
Most tools have clearly distinct purposes: cost estimation, usage lookup, parameter discovery, content operations, catalog operations, and payment retry. The three content-generation tools share very similar generic descriptions, but their names still distinguish title enrichment, market research, and content generation.
All tool names use a consistent snake_case verb_noun pattern, such as estimate_cost, get_usage, enrich_title, and retry_payment. The few longer names like get_catalog_status remain predictable and readable.
Nine tools is well scoped for this API surface, covering estimation, account state, schema discovery, content workflows, catalog workflows, and payment retry. Each tool appears to earn its place without obvious filler.
The set covers the main lifecycle: estimate cost, inspect usage and parameters, run title/research/content operations, submit and track catalog enrichment, and retry payment. Minor gaps may exist around retrieving status or results for non-catalog content requests, but core workflows are represented.
Maintenance
Related MCP Connectors
Clean, matched, AI-ready product data from any website
Domain & brand intelligence: company enrichment, tech stack detection, brand research.
Writes and SEO-scores product listings from the seller's facts, for 7 marketplaces in 10 languages.
Live Amazon product, review, search, deal, offer/stock and seller data across 23 marketplaces.
Related MCP Servers
- AlicenseBqualityCmaintenanceProvides 19 B2B data enrichment tools for finding and validating emails, searching company profiles, analyzing job postings, discovering employee information, and researching advertising campaigns through the LeadMagic API.1929 npm7MIT
- AlicenseAqualityDmaintenanceKlarna-style product discovery for AI shopping agents. Makes product catalogs machine-readable so AI agents can search, compare, and purchase products programmatically.637 PyPIMIT
- AlicenseNot gradedqualityCmaintenanceEnables product enrichment with sentiment analysis, category mapping, and attribute extraction; provides tools to fetch top products and cluster summaries.MIT
- FlicenseNot gradedqualityDmaintenanceNormalizes messy product catalog attributes and maps products to taxonomies like Google and Shopify, with retrieval-grounded category classification.2-