Swervie

Swervie MCP server

Community Swervie
Updated

Swervie MCP server: one scored answer per product, across every retailer, for AI agents. Remote Streamable HTTP, no auth.

Swervie MCP server

One scored answer per product, across every retailer, for AI agents. Remote MCP server · Streamable HTTP · no auth · no API key · free

Developer docs · Tool reference · Connect a client · Response shapes · Data handling

https://us-central1-swervie-prod.cloudfunctions.net/mcp

Swervie is a cross-retailer product catalog built for assistants. Ask a normalshopping question and get one consolidated, scored answer per product: liveprices, comparable specifications, review-derived ratings with verbatim evidencequotes, and a buy link for every seller. Instead of eight duplicate listings anda stale listicle.

  • Consolidated. Listings of the same product across retailers collapse intoone row. White-label clones sold under different brands are grouped too.
  • Structured. Every category has its own spec schema; specs are extractedfrom manufacturer pages, retailer listings and spec tables, with provenance.
  • Scored. Each product carries a quality score and a value score, plusreview-derived sub-ratings backed by quotes from real reviews.
  • Deterministic and compact. No LLM in the serving path. Answers areprecomputed pipeline output: ~4k tokens, sub-second for a full shelf.
  • Interactive in hosts that support MCP Apps. Results render as a shelfwith sort controls, seller-by-seller offers and the review evidence(io.modelcontextprotocol/ui extension); plain JSON everywhere else.

Coverage today: 300+ categories, from robot vacuums and televisions toespresso machines, mattresses and hiking boots, across hundreds of sellers.The live list is at swervie.xyz/llms.txt.

Quick start

Claude Code

claude mcp add --transport http swervie https://us-central1-swervie-prod.cloudfunctions.net/mcp

Cursor, Windsurf, VS Code, Claude Desktop, ChatGPT, Gemini CLI and others:see docs/clients.md. Every config is the same one line, theURL above, with no headers and no secrets.

curl

MCP=https://us-central1-swervie-prod.cloudfunctions.net/mcp
H=(-H 'Content-Type: application/json' -H 'Accept: application/json, text/event-stream')

# Handshake: server info, capabilities, agent instructions
curl -s -X POST $MCP "${H[@]}" -d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2025-06-18","capabilities":{},"clientInfo":{"name":"curl","version":"1.0"}}}'

# A whole shopping answer in one call
curl -s -X POST $MCP "${H[@]}" -d '{"jsonrpc":"2.0","id":2,"method":"tools/call","params":{"name":"search_products","arguments":{"query":"65 inch OLED TV under $1500","limit":3}}}'

More in examples/: a Node client using the official SDK, a Pythonclient, and the curl script above.

Tools

Tool Kind What it does
search_products read The one-call path. Resolves a free-text query to a named product, an "A vs B" comparison, or a category shelf with constraints as filters and wishes as ranking preferences. Returns an interpretation receipt.
lookup_product read One product by name, ASIN, UPC/EAN/GTIN, or brand + model. Full detail, every seller offer, and its rank on its shelf.
find_category read The matching category and its filter contract: every filterable spec with type, units, allowed values and importance, plus price bands.
browse_category read A category's ranked shelf with explicit filters (hard) and preferences (soft), sorted by value, quality or price.
get_products read Full detail for up to 25 ids: all specs with provenance, review sub-ratings with evidence quotes, every offer.
record_selection write, non-destructive Which product the user considered or chose. Appends one event; changes nothing else.
submit_feedback write, non-destructive A problem report for human review: wrong spec, stale price, bad match, missed duplicate, missing product.

All tools carry MCP annotations (title, readOnlyHint, destructiveHint,openWorldHint), so hosts can run the reads without a confirmation prompt.Full input schemas and descriptions, generated from the live server, are indocs/tools.md.

A typical conversation

  1. "Find me a 65-inch OLED TV under $1,500" → search_products → ranked shelfwith prices, buy links and the commission disclosure.
  2. "Which of those is better for gaming?" → get_products → refresh rate, VRR,HDMI 2.1, and what reviewers say about input lag.
  3. "I'll take the Samsung" → record_selection → improves later rankings.

The server ships instructions on initialize that teach an agent this flow,so a client needs no prompt engineering.

Personalization without PII

Every tool accepts an optional user_signature: an opaque, stable, pseudonymousid that makes rankings consistent for a returning user across agents andsessions. The recommended recipe:

user_signature = "sv1-" + sha256hex("swervie/v1|" + id).slice(0, 24)

where id is the user's most stable identifier, normalized (lowercase,trimmed; email preferred, with +tags stripped). Never send the rawidentifier. If none is available, use "sv1r-" + a UUID persisted in theagent's own memory for that user, or omit the field. The server hashes any rawemail or phone number that arrives by mistake before anything is stored.

Limits, affiliate links, and privacy

  • Rate limits. Anonymous tier: 60 requests/min and 2,000/day per client(signature, session, or IP), with a higher per-IP ceiling sized for sharedagent-platform egress. Partner keys with their own quota are available forplatforms: contact [email protected].
  • Buy links are affiliate links. Every response that carries buy_urlalso carries a disclosure string worded for the shopper. Commissions neveraffect rankings, scores, or which products are shown.
  • No account, no checkout, no payments. Purchasing hands off to theretailer's own page.
  • What is stored. Tool name, query, category and product ids per call, thesignature when present, and rate-limit counters. Nothing the user owns.

Details in docs/data-handling.md. Policies:privacy · terms ·disclosure.

What is in this repository

This repository is the public home of the Swervie MCP server: documentation,the registry manifest (server.json), client configuration, examples, andbrand assets for directories. The serving code runs as a Google Cloud Functioninside Swervie's main monorepo and is not published here. The hosted endpointis the supported way to use it. Issues and feature requests are welcome.

Server version is reported by initialize (serverInfo.version). Protocolversion 2025-06-18. Transport: Streamable HTTP, stateless.

License

Documentation, examples and assets in this repository are MIT licensed(© Plum Applications, Inc.). Use of the hosted service is governed by theterms.

MCP Server · Populars

MCP Server · New