Connect your agent to Comparare
When a user says “I need one product for CRM and invoicing; the vendor should have 1,000+ employees and 10k active users”, your agent finds the right endpoint and gets a structured answer in seconds. Reading is free and needs no key.
1. MCP (Claude, ChatGPT, Cursor, VS Code…)
Our Model Context Protocol endpoint uses Streamable HTTP and needs no sign-in. Server URL:
claude.ai, Claude Desktop and Claude mobile share the same connectors.
- Open Settings → Connectors.
- Click Add custom connector.
- Name:
Comparare, remote MCP server URL:https://comparare.echoenes.dev/mcp. Leave the OAuth fields under Advanced settings empty. - Click Add. In a chat, turn Comparare on from the + → Connectors menu and ask your question.
Team and Enterprise: an organization owner first adds the connector under Admin settings → Connectors; members then enable it in their own settings.
Custom MCP servers are added through developer mode (Plus, Pro, Business, Enterprise and Edu, on the web).
- Open Settings → Apps → Advanced settings and turn on Developer mode.
- Click Create app (called "Create connector" in some versions).
- Name:
Comparare, MCP server URL:https://comparare.echoenes.dev/mcp, authentication: No authentication. Confirm that you trust the app and save. - In a new chat open +, pick Developer mode and enable Comparare.
Business and Enterprise workspaces: an admin may need to allow developer mode first. Without developer mode, use a Custom GPT with the OpenAPI definition below.
Claude Code:
claude mcp add --transport http comparare https://comparare.echoenes.dev/mcp
Cursor, VS Code, Windsurf and other clients (mcp.json):
{
"mcpServers": {
"comparare": {
"type": "http",
"url": "https://comparare.echoenes.dev/mcp"
}
}
}
Cloud clients (claude.ai, ChatGPT) can only reach a public HTTPS address; local development URLs will not work there.
Tools
search_productslist_categoriesget_productget_vendorcompare_productsfind_alternativesEvery tool accepts an optional "language" argument (en, tr) for localized text.
2. REST API + OpenAPI (Custom GPT Actions, LangChain, n8n…)
Import the OpenAPI 3.1 definition: https://comparare.echoenes.dev/api/v1/openapi.json
# Turn the user's sentence into filters GET https://comparare.echoenes.dev/api/v1/parse?q=one product for crm and invoicing, at least 1000 employees, 10k active users # Or build the filters yourself (categories are AND-ed) GET https://comparare.echoenes.dev/api/v1/products?categories=crm,invoicing&min_employees=1000&min_active_users=10000&include_unknown_metrics=true # Only Deloitte alliance partners GET https://comparare.echoenes.dev/api/v1/products?categories=supply-chain-planning&deloitte_alliance=true # Turkish text in responses GET https://comparare.echoenes.dev/api/v1/products/odoo?lang=tr # Comparison GET https://comparare.echoenes.dev/api/v1/compare?products=hubspot-crm,salesforce-sales-cloud
Unknown metrics are null — your agent should not guess them. Every product carries deloitte_alliance. /compare also returns pros and cons per product.
3. llms.txt
Summary for agents browsing the web: /llms.txt · full catalog: /llms-full.txt. Product pages also include schema.org SoftwareApplication JSON-LD.
Sample system prompt (Custom GPT / agent)
When the user describes a software need, call parseQuery with the user's sentence, then searchProducts with the returned filters (add include_unknown_metrics=true if nothing matches). Recommend the best 3 products with price, vendor size and user count; if a value is null say "unknown" instead of guessing. Mention when a product is a Deloitte alliance partner (deloitte_alliance). Give the "url" link for each product. If the user asks for a comparison, call compareProducts and summarise its pros and cons.