For agents & developers
Everything an agent can call,
in one place.
This site exposes a live MCP tool server backed by a real Jennifer Furniture deployment — plus every machine-readable surface a crawler, assistant, or script needs. No API key required for the public tool surface.
Endpoints & files
| POST | /mcp | JSON-RPC 2.0 MCP endpoint — initialize, tools/list, tools/call, resources/list, resources/read |
| GET | /.well-known/mcp | MCP server card — endpoint, transport, tool list |
| GET | /.well-known/ai-catalog.json | AI catalog manifest of every agent-facing resource |
| GET | /openapi.json | OpenAPI 3.1 description of /mcp |
| GET | /llms.txt | Site manifest |
| GET | /llms-full.txt | Entire site content in one file (also served as Markdown via Accept: text/markdown on /) |
| GET | /AGENTS.md | Agent capabilities and boundaries |
Call the storefront tools
11 public catalog tools run against the live Jennifer Furniture deployment — product search, variant availability, room-fit math, floorplan SVGs, room bundles, AR links, checkout handoff. Admin tools are gated upstream and return -32601.
POST https://tendrils.io/mcp
Content-Type: application/json
{"jsonrpc":"2.0","id":1,"method":"initialize",
"params":{"protocolVersion":"2024-11-05","capabilities":{},
"clientInfo":{"name":"my-agent","version":"1.0"}}}{"jsonrpc":"2.0","id":2,"method":"tools/list","params":{}}
{"jsonrpc":"2.0","id":3,"method":"tools/call","params":{
"name":"find_products_by_constraints",
"arguments":{"query":"sleeper sofa under $1500","limit":4}}}Conventions
- Rate limit: 60 requests/minute per IP. Responses carry
RateLimit-*headers; a 429 returnsRetry-After. - Errors: JSON-RPC error objects with machine-readable codes (-32700 parse, -32600 invalid, -32601 not available, -32602 unknown resource, -32603 upstream, -32000 rate limit).
- Versioning: the endpoint is versioned by the MCP
protocolVersionin the initialize handshake; breaking changes ship behind a new version, never silently. - Content negotiation: send
Accept: text/markdownon the homepage for the Markdown mirror. - Crawl policy: robots.txt declares
search=yes, ai-input=yes, ai-train=no.
Building an agentic surface for your own store is what Tendrils does for clients — talk to us.