Catalog is an MCP server with 11 tools an agent can call, among them build_bom, check_compatibility, compare_parts. It answers at https://partsgraph.ai/api/mcp over streamable HTTP, and it listed its tools to an anonymous client, so no sign-in was needed for that. Our probes reached it 10 of 10 times since 4 September 2026; the last check, on 17 October 2026, succeeded.
Search real parts with datasheet-provenance specs, check compatibility and compose priced BOMs.
Description as the publisher wrote it.
| Tool | What it does |
|---|---|
build_bom | Composes a complete BOM for a reference build — quantities, dated indicative prices, per-line rationale, a compatibility verdict for the chosen set, and a CSV export. Override any role with choices. Inputs: build_id, choices, project_context |
check_compatibility | Evaluates a part set against the build's engineering constraints — supply windows, stall vs peak current, shaft/wheel fit, interfaces, solar chain — with a reasoned verdict per check. Every number cited is a corroborated attribute. Inputs: part_ids, build_id, project_context |
compare_parts | Aligned attribute-by-attribute comparison of two catalogue parts. Absence means the sources did not state a value — never that the part lacks the property. Inputs: a, b, project_context |
fetch | Fetch the full provenance-backed record for a search result id as a citable document. Alias of get_part shaped to the search/fetch connector contract. Inputs: id |
get_part | The complete canonical record for one part: every attribute with its provenance (source, page, verbatim quote, file hash), vendors with dated indicative prices, and datasheet links. Inputs: id, project_context |
get_provenance | The receipt behind a stated value: the verbatim source quote, the page, the extraction method, and the sha-256 of the exact file it was read from — plus a ready-made citation line. Omit code for every attribute's receipt. Inputs: id, code, project_context |
list_builds | The catalogue's reference builds (robot, weather station, plant waterer) with their roles and candidate parts — the starting point for build_bom. Inputs: project_context |
list_categories | Every category in the catalogue with its part count. Inputs: project_context |
search | Search the Partsgraph demo catalogue. Returns result ids for fetch. Alias of search_parts shaped to the search/fetch connector contract. Inputs: query |
search_parts | Full-text search over the demo catalogue's provenance-backed part records. Filter by category, build or role. Every attribute returned was corroborated against the manufacturer datasheet or vendor page. Inputs: query, category, build, role, limit, project_context |
view_bom | Composes the BOM for a reference build and renders it as an interactive view in clients that support MCP Apps. Same data as build_bom. Inputs: build_id, choices |
Every block below points at https://partsgraph.ai/api/mcp. Pick your client and copy.
claude mcp add --transport http catalog https://partsgraph.ai/api/mcp
{
"mcpServers": {
"catalog": {
"command": "npx",
"args": ["-y", "mcp-remote", "https://partsgraph.ai/api/mcp"]
}
}
}{
"mcpServers": {
"catalog": {
"url": "https://partsgraph.ai/api/mcp"
}
}
}{
"servers": {
"catalog": {
"type": "http",
"url": "https://partsgraph.ai/api/mcp"
}
}
}{
"context_servers": {
"catalog": {
"url": "https://partsgraph.ai/api/mcp"
}
}
}{
"mcpServers": {
"catalog": {
"serverUrl": "https://partsgraph.ai/api/mcp"
}
}
}{
"mcpServers": {
"catalog": {
"url": "https://partsgraph.ai/api/mcp"
}
}
}name: Catalog
version: 0.0.1
schema: v1
mcpServers:
- name: Catalog
type: streamable-http
url: https://partsgraph.ai/api/mcp{
"mcpServers": {
"catalog": {
"httpUrl": "https://partsgraph.ai/api/mcp"
}
}
}{
"mcpServers": {
"catalog": {
"type": "http",
"url": "https://partsgraph.ai/api/mcp"
}
}
}{
"mcpServers": {
"catalog": {
"url": "https://partsgraph.ai/api/mcp"
}
}
}goose://extension?url=https%3A%2F%2Fpartsgraph.ai%2Fapi%2Fmcp&type=streamable_http&id=catalog&name=Catalog
It answered our last check on 17 October 2026 in 181 ms. Over 10 checks since 4 September 2026 it answered 10. Status here comes from Neuronto's own probes, not from the publisher.
Yes. It is published by partsgraph.ai and served from partsgraph.ai, the vendor's own domain: https://partsgraph.ai/api/mcp.
11, as returned by the server's own tools/list on 8 September 2026: build_bom, check_compatibility, compare_parts, fetch, get_part, get_provenance, list_builds, list_categories, and 3 more.
It answered our last check on 17 October 2026 in 181 ms. Over 10 checks since 4 September 2026 it answered 10.
Put the server under mcpServers in ~/.cursor/mcp.json with the URL https://partsgraph.ai/api/mcp. The exact JSON is above, with the same setup for Claude Code, Claude Desktop, VS Code, Windsurf, Zed, Gemini CLI and the rest.
It listed its tools to an anonymous client, so reading the tool list needs no key. Individual tools can still ask for credentials in their own inputs.
Publisher: ai.partsgraph. Endpoint host: partsgraph.ai. ARD identifier: urn:air:ai.partsgraph:mcp:catalog.
First indexed 31 August 2026. The tools are exactly what the server returned; nothing on this page
is inferred from its name. Ranking in Neuronto search is never for sale.