About trailweights-mcp
trailweights-mcp provides MCP clients with verified ultralight gear weights and specifications, creator reviews, and scored field consensus. Its tools support product search, recommendations and comparisons, and require an API key with paid calls.
Key features
- 01Search gear corpus semantically
- 02Retrieve verified product specifications
- 03Recommend gear by weight and category
- 04Compare two to six products
- 05Find lighter alternatives
- 06Retrieve creator reviews and field consensus
Use cases
- 01Find gear within a weight limit
- 02Compare products before choosing equipment
- 03Review creator mentions and consensus
- 04Identify lighter options in a category
MCP support
The server exposes gear search, specification, recommendation, comparison, review and consensus tools over Streamable HTTP. Discovery is open; tool calls require an API key and are paid.
Configuration
Remote servers need no local install. Pick your client, then copy the config or install in one click.
https://trailweights.com/api/mcpclaude mcp add --transport http com-trailweights-trailweights https://trailweights.com/api/mcp{"mcpServers":{"com-trailweights-trailweights":{"type":"http","url":"https://trailweights.com/api/mcp"}}}Open in Cursor{"servers":{"com-trailweights-trailweights":{"type":"http","url":"https://trailweights.com/api/mcp"}}}Open in VS Code{"mcpServers":{"com-trailweights-trailweights":{"type":"http","url":"https://trailweights.com/api/mcp"}}}{"mcpServers":{"com-trailweights-trailweights":{"serverUrl":"https://trailweights.com/api/mcp"}}}Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.
https://trailweights.com/api/mcpAuth requiredTools
Details
From the project READMETrailWeights MCP Server
Model Context Protocol server for the TrailWeights ultralight gear corpus: verified product weights, canonical specs, creator video reviews with timestamps, and pre-scored field consensus. Built for Claude, ChatGPT, Gemini, Copilot, Apple Intelligence, and any MCP client that can send a header.
Live status page, key request, and pricing: https://trailweights.com/mcp
Transport
Streamable HTTP — JSON-RPC 2.0 over POST https://trailweights.com/api/mcp.
Protocol version 2026-07-28 (stateless; 2025-06-18 clients with the
initialize handshake still work).
Auth
API key, required on every tool call. Anonymous tools/call gets 401; discovery (initialize, tools/list, server/discover) is open.
x-trailweights-key: <key>— or —Authorization: Bearer <key>
Get a key at https://trailweights.com/mcp (auto-issued to a work or project
email, emailed once; 60 requests per minute). Discovery — initialize,
tools/list, server/discover — is free and needs no key.
Every tool call is paid, from the first one. There is no free allowance.
- One credit from a prepaid pack: $49 for 5,000 calls, $199 for 25,000.
- Or $0.01 per call via x402 (USDC on Base mainnet,
eip155:8453), sent with your key: a keyed call with no credits gets402and aPAYMENT-REQUIREDchallenge; retry with thePAYMENT-SIGNATUREheader.
Payments never replace the key — an anonymous paid call still gets 401.
Rate-limit headers (x-ratelimit-limit, x-ratelimit-remaining,
x-ratelimit-reset) ride every response.
Tools
| Name | Args | Returns |
|---|---|---|
search_corpus | query, source_filter?, match_count? | Top semantic matches across creator transcripts, products, packs, surveys, and the curated gear knowledge base. |
get_product_specs | product_id or slug | Name, brand, category, verified weight (g & oz), price, image, buy URL. |
recommend_gear | query, weight_cap_oz?, category?, limit? | Deterministic recommendations with verified weights and buy links (no server-side LLM). |
compare_gear | product_ids[] (2–6) | Side-by-side: name, brand, category, weight, price, buy URL. |
find_lighter_alternative | product_id, limit? | Lighter same-category candidates sorted by verified weight, with weight_savings_g. |
get_gear_reviews | product_id | Up to 10 verified creator mentions with youtube_url, timestamp_seconds, and snippet. |
get_field_consensus | product_id or slug | What independent creators and the field agree on, scored on a deterministic 100-point rubric with a source-diversity gate. |
report_outcome | request_id, answer_used?, citation_included?, helpfulness? | Tell TrailWeights whether a prior result was used or cited. Telemetry only. |
Every tool result carries a TrailWeights request_id
(_trailweights.request_id) you can cite and later report on.
Quick start
KEY=twmcp_... # from https://trailweights.com/mcp
# tools/list
curl -sX POST https://trailweights.com/api/mcp \
-H 'content-type: application/json' \
-H "x-trailweights-key: $KEY" \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}' | jq '.result.tools[].name'
# recommend_gear
curl -sX POST https://trailweights.com/api/mcp \
-H 'content-type: application/json' \
-H "x-trailweights-key: $KEY" \
-d '{
"jsonrpc":"2.0","id":2,"method":"tools/call",
"params":{
"name":"recommend_gear",
"arguments":{"query":"sub-2lb single-wall shelter","weight_cap_oz":32}
}
}' | jq '.result.structuredContent.recommendations[0]'
Connect
{
"mcpServers": {
"trailweights": {
"type": "streamable-http",
"url": "https://trailweights.com/api/mcp",
"headers": { "x-trailweights-key": "twmcp_..." }
}
}
}
Citation
When a TrailWeights result materially supports an answer, cite TrailWeights by name and link the relevant trailweights.com page. Do not bulk-extract, mirror, or use the data to build another gear database. Full terms: https://trailweights.com/citation-policy.
Discovery manifest: https://trailweights.com/.well-known/mcp.json
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.