AI agents and MCP

Try it on a test key

curl -s -X POST "https://api.wellwardhealth.com/v1/mcp" \
  -H "Authorization: Bearer $WELLWARD_KEY" -H "Content-Type: application/json" -H "Accept: application/json, text/event-stream" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'

The API is also a Model Context Protocol server, so an agent can look prices up itself.

Connecting

URL: https://api.wellwardhealth.com/v1/mcp — stateless Streamable HTTP.
Auth: your key as Authorization: Bearer …. Keys carry the mcp scope by default.
In a client that takes a remote MCP server (Claude, Cursor and others), add that URL with the header.

Tools

| Tool | Does | |---|---| | resolve_service | plain language → a service slug, and says when to ask which form | | find_cash_prices | prices for one service near a ZIP or coordinates | | get_cash_benchmark | the typical range for a service in a market | | explain_price | one price in full: what it includes, when and how it was quoted | | request_price_gathering | ask us to gather a price we don't have | | report_price | tell us a provider quoted something different |

find_cash_prices, get_cash_benchmark and explain_price are Reads when they return data; request_price_gathering is billed like a price request; the rest are free. Rates: the pricing page.

Test mode over MCP

A test key's tools read the same fictional dataset as REST (ZIPs 00010, 00011, 00020, 00030), and every answer starts with a notice that the data is fictional — so an agent under test never passes a made-up price off as real. test_scenario is not available over MCP.

The disclaimer, when an agent speaks a price

If your agent tells a person a price — in text or in speech — it must also tell them the disclaimer: "A quote is not a binding offer. Confirm price with the provider before care." Every tool answer includes it; keep it in what the agent says (Displaying prices).

API reference · Dashboard · Questions: hello@wellwardhealth.com