OpenRemedies
This server has a remotely hosted endpoint. Open its connector listing for connection details, authentication requirements, and setup instructions:
OpenRemedies
OpenRemedies is a lookup of sourced treatment records (approved, off-label, supplements, in development, recruiting). It never recommends treatment, diagnoses, doses, or declares trial eligibility. Every record carries a citation; empty results mean no records in covered sources, not that no treatments exist. Built on pipeworx.
Part of Pipeworx — an MCP gateway connecting AI agents to 1856+ live data sources. This is an independent, unofficial integration — not affiliated with, endorsed by, or published by the upstream provider.
(That paragraph is the medical-content posture from the OpenRemedies MCP
endpoint's initialize.instructions, preserved verbatim per fleet #1392.
Keep it word-for-word in any public catalog entry for this pack.)
This pack proxies to the OpenRemedies MCP endpoint (Bruce's free
treatment-lookup product, repo b-gutman/openremedies). It is keyless,
read-only public data; OpenRemedies applies its own rate limiting at its
Worker. Every response carries meta with source citations, per-source
freshness, data_version, and coverage notes.
Tools
Tool | What it returns |
| Condition name/lay term/misspelling → condition concepts (candidates, never merged) |
| Drug/herb/supplement name → substance records with labeled salt/form family |
| Five fixed sections for a condition: approved, documented off-label, herbs/supplements, in development, recruiting |
| Approvals by jurisdiction, verbatim indications, evidence assertions |
| Registry trials by condition/substance/status/phase/country |
| Broader/subtype/sibling condition concepts, each labeled by its structural relationship (type of / subtype / related type). Derived from the disease hierarchy — not an assertion of clinical relatedness |
| One full evidence assertion with verbatim finding and provenance |
| What is and is NOT covered: per-source freshness, jurisdiction matrix, evidence counts |
Typical flow: resolve_condition("migraine") → condition_id →
get_condition_options(condition_id); or resolve_substance("sumatriptan")
→ substance_id → get_substance_uses(substance_id).
get_related_concepts takes the same condition_id and answers "what else
should I look at" — it returns an empty list for concepts with no hierarchy
neighbours carrying records (e.g. psoriasis, atopic eczema), which means
no neighbours in the covered hierarchy, not that none exist.
Related MCP server: meps
Listed; origin flipped to the apex (fleet #3511)
Bruce lifted the unlisted hold for search and routing on 2026-10-21. The pack
is in the catalog, in search_packs and on routing menus. On 2026-10-26 the
OpenRemedies apex began answering (it had returned 522 that morning), and
ORIGIN in src/index.ts now points at it. On the same calls the apex and the
former staging host returned identical tools/list and get_coverage bodies,
so the apex is currently the same deployment. It still reports a staging
version, so every response keeps its deployment notice until the upstream
reports a production version. Adding a registry-api /uptime row remains a
launch item for the OpenRemedies PM.
The tool definitions in src/index.ts are copied from the endpoint's
tools/list. If OpenRemedies changes its tool surface, update them to match;
nothing syncs them automatically. Coverage figures (sources, jurisdictions)
come from get_coverage; never hard-code them in a description.
Environment disclosure (fleet #3660)
Every response carries a top-level deployment object (environment,
upstream_version, upstream_host, notice). It is derived, not written by
hand: the pack reads the upstream's own reported version and the host it
called. A staging version or host yields non-production with a notice that
promises no uptime or freshness; a clean version on a clean host yields
production; an unreadable version yields unknown. When ORIGIN flips and
the upstream reports a production version, the non-production wording stops by
itself. src/index.test.ts fails if a production-looking version produces the
non-production wording.
Quick Start
Add to your MCP client (Claude Desktop, Cursor, Windsurf, etc.):
{
"mcpServers": {
"openremedies": {
"url": "https://gateway.pipeworx.io/openremedies/mcp"
}
}
}What this endpoint actually serves
tools/list at https://gateway.pipeworx.io/openremedies/mcp returns the tools in the table
above plus the shared Pipeworx meta-tools — ask_pipeworx,
discover_tools, search_within, remember/recall and the rest of the
gateway-wide set. So the tool count you see is larger than this table: a
single-pack endpoint currently lists roughly 30 shared tools alongside the
pack's own. The connection's initialize response states its exact scope, and
is the authoritative answer for a given day.
This is deliberate, not multiplexing by accident. The meta-tools are what let a
scoped connection answer a question this pack does not cover — via
ask_pipeworx, which routes across the whole catalog — without you adding a
second MCP server. There is currently no way to mount a pack endpoint without
them; if the extra schemas cost you more context than the routing is worth,
connect to the full gateway once rather than to several pack endpoints.
Or connect to the full Pipeworx gateway to get every pack's tools listed directly, instead of just this one's:
{
"mcpServers": {
"pipeworx": {
"url": "https://gateway.pipeworx.io/mcp"
}
}
}Both URLs reach the same gateway and the same 1856+ data sources. The
only difference is which pack's tools are listed directly; ask_pipeworx
reaches all of them from either one.
No MCP client? Call it over HTTP
curl -X POST https://gateway.pipeworx.io/v1/tools/resolve_condition \
-H 'Content-Type: application/json' \
-d '{"q":"migraine","limit":3}'No account needed for the first calls. Inspect any tool: GET https://gateway.pipeworx.io/v1/tools/resolve_condition. Find one: POST https://gateway.pipeworx.io/v1/tools/search_packs with {"query":"..."}.
Run it from npm
@pipeworx/mcp-openremedies is a local stdio bridge to the Pipeworx gateway: your MCP client
starts it with npx, and it forwards every call to the same
https://gateway.pipeworx.io/openremedies/mcp endpoint shown above. It needs network access;
there is no offline mode.
{
"mcpServers": {
"openremedies": {
"command": "npx",
"args": ["-y", "@pipeworx/mcp-openremedies"]
}
}
}Without a key it calls anonymously, capped at 50 calls a day. A free account
raises that to 200 a day: get a key at https://pipeworx.io/signup and add
"env": { "PIPEWORX_API_KEY": "<your key>" } to the server entry.
The key is sent as a Bearer token, the same way a hosted connection sends it,
so calls are metered and billed exactly as they are there.
It lists only this pack's tools by default. Set PIPEWORX_META_TOOLS=1 to
also list ask_pipeworx and the other gateway-wide tools.
Behind an HTTPS proxy: Node's built-in fetch does not read HTTPS_PROXY by
default. Newer Node releases honour it when NODE_USE_ENV_PROXY=1 is set; check
your Node version's documentation.
Using with ask_pipeworx
Instead of calling tools directly, you can ask questions in plain English — this works on the pack endpoint above as well as on the full gateway:
ask_pipeworx({ question: "your question about Openremedies data" })The gateway picks the right tool and fills the arguments automatically.
More
License
MIT
This server cannot be deployed
Maintenance
Related MCP Connectors
Condition-aware ingredient and product checks for agents, with evidence tiers and citations.
Agent research API for 15 disease areas, plus a T1D cure lab, sourced evidence ledger and models.
Clinical trials, RxNorm drug lookups, FDA drug and device data, PubMed and NHS organisations.
Physician-reviewed medical opinions and prescriptions for AI agents.
Related MCP Servers
- FlicenseNot gradedqualityDmaintenanceEnables searching and retrieving information from the ClinicalTrials.gov database of over 400,000 clinical studies, including trial details, eligibility criteria, locations, and results across 220+ countries.5-
- AlicenseNot gradedqualityBmaintenanceQueries US prescription drug use by medical condition from AHRQ's Medical Expenditure Panel Survey — which drugs Americans take for a given condition, what a drug is prescribed for, and national volume, spending and out-of-pocket share for individual drugs or the top drugs of a year. Every response carries sample sizes and reliability flags so estimates are treated as rough magnitude and ranking rather than precise counts.459 npmMIT
- AlicenseNot gradedqualityCmaintenanceEnables AI assistants and MCP clients to query FDA drug labels, FAERS adverse-event reports, recalls, and label-documented drug interactions, returning structured, cited results.MIT
- AlicenseNot gradedqualityCmaintenanceEnables AI agents to resolve genes, diseases, chemicals, variants and species to normalized ids, search ~36M PubMed abstracts and PMC full texts by free text or entity, and retrieve the machine-extracted relations between them with the supporting sentences and PMIDs. Supports auditing individual relations with full evidence passages and pulling per-article entity annotations with character offsets.255 npmMIT