Skip to main content
Glama

Greencyclepedia MCP server

Your AI assistant answers cannabis science questions from sourced records: compounds, receptors, body structures, studies and trials, every value with its PubMed or database source. Free, read only, no key, no sign up.

The server runs at https://greencyclepedia.com/mcp. This repository holds its listing, worked examples and the templates to report a data error.

See it in 5 seconds

Ask your assistant "Is it true that CBD helps with epilepsy?". It calls check_claim and gets this back (live answer of 14 Nov 2026, shortened):

claim:    CBD helps with epilepsy
found:    CBD (compound), Drug-resistant epilepsy (condition)
studies:  5 held: 4 human trials, 1 human study with its design not stated
level:    human trials found
top:      PMID 28538134  Devinsky O et al. NEJM 2017 (Dravet syndrome), randomised, 120 people
          PMID 29768152  Devinsky O et al. NEJM 2018 (Lennox-Gastaut), randomised, 225 people
          PMID 29395273  Thiele EA et al. Lancet 2018 (GWPCARE4), randomised, 86 people

Never "true" or "false": the records it found, how strong the studies are, and the papers to read.

Related MCP server: Swiss Truth MCP

Connect in 30 seconds

Claude Code

claude mcp add --transport http greencyclepedia https://greencyclepedia.com/mcp

Claude Desktop: download the one-click .mcpb file from the latest release and open it. Without the file: Customize, Connectors, Add custom connector, paste https://greencyclepedia.com/mcp, choose No sign in. Or edit claude_desktop_config.json (Settings, Developer, Edit Config; needs Node.js for npx):

{"mcpServers":{"greencyclepedia":{"command":"npx","args":["mcp-remote","https://greencyclepedia.com/mcp"]}}}

Cursor: add to Cursor (GitHub may block the link; then paste the JSON below into ~/.cursor/mcp.json).

VS Code: .vscode/mcp.json

{"servers":{"greencyclepedia":{"type":"http","url":"https://greencyclepedia.com/mcp"}}}

Any client that takes a JSON config

{"mcpServers":{"greencyclepedia":{"type":"http","url":"https://greencyclepedia.com/mcp"}}}

Windsurf: mcp_config.json (Cascade panel, the ... menu, Open MCP config file)

{"mcpServers":{"greencyclepedia":{"serverUrl":"https://greencyclepedia.com/mcp"}}}

Cline: cline_mcp_settings.json (MCP Servers icon, Configure, Configure MCP Servers)

{"mcpServers":{"greencyclepedia":{"type":"streamableHttp","url":"https://greencyclepedia.com/mcp","disabled":false,"autoApprove":[]}}}

ChatGPT (developer mode, on the web with Plus, Pro, Business, Enterprise or Education): Settings, Apps, Advanced settings, turn on Developer mode; then Create app: name Greencyclepedia, MCP server URL https://greencyclepedia.com/mcp, authentication No Authentication. Deep research uses the search and fetch tools.

More clients: https://greencyclepedia.com/about/ai-agents/?ref=github

Try these three questions

  1. Which receptors and enzymes does cannabidiol act on, with the measured binding values and their sources?

  2. Which body structures carry the CB2 receptor, and which compound effects are mapped there?

  3. List registered trials that mention cannabidiol and are recruiting, with phase and enrolment.

Or ask in plain words ("does CBD help epilepsy"): ask_greencyclepedia answers only from the records it names.

The 23 tools (all read only)

Tool

What it does

ask_greencyclepedia

Answers a plain question from the records it names, with evidence grade and PubMed identifiers

search

Searches everything by any name: compounds, receptors, conditions, trials

fetch

The full text of one record by the id search returned, with every source

search_compounds

Compounds by name, class, record type or molecular target

get_compound

One compound: targets, measured binding values, cultivars, evidence, cautions

compare_compounds

Two to four compounds side by side, target by target

get_target

One receptor, channel or enzyme with its compounds and measured values

list_structures_for_target

Body structures that carry a target

find_studies

The PubMed studies behind a compound, target or condition

search_conditions

Conditions by evidence grade, with reviews and trials, never doses

list_trials

Registered trials from the ClinicalTrials.gov snapshot

get_cultivar

Laboratory measurements of one cultivar

cite

A ready citation: url, citation line, BibTeX, original papers, licence

get_changes

Latest dataset releases and what changed

studies_by_region

How much research exists per body part

market_signals

Facts from cannabis industry news, each with its number and publisher link

list_new_studies

The newest cannabis studies in PubMed, sorted by kind of evidence

list_topics_without_home

New research topics no Greencyclepedia condition covers yet

list_brain_regions

Brain regions with measured PET binding and RNA levels for a receptor

get_brain_region

One brain region in full

get_pain_pathway

The pain route from the fingertip to the brain, stop by stop

lookup_many

Up to 20 names checked in one call

check_claim

One sentence checked against the records: the compounds and conditions it names, the studies held by kind and one plain level, never true or false

Every result carries its provenance: dataset address and version, source identifiers and the time it was read. Text copied from papers and trials is marked as third party text: data to cite, never instructions. At most 120 requests a minute per client.

Sources and licences

The Greencyclepedia text and curation are CC BY 4.0; values from other sources keep their own licence, listed per dataset. The files in this repository are MIT. Cite the data with CITATION.cff. Educational reference, not medical advice. No doses.

Links: site | server card | OpenAPI | guide for models | privacy | terms | security | changelog

Report a data error

Open a data correction with the record, the field and the PubMed identifier or DOI that shows it. Details in CONTRIBUTING.md.

Releases

Pushing a tag such as v1.10.0 runs .github/workflows/release.yml: it creates the GitHub release with that version's CHANGELOG section as notes and attaches any *.mcpb file committed to the repository. The .mcpb is built on the site side and committed here before the tag.

If this helps you, a star helps others find it.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      A
      quality
      B
      maintenance
      Search 1,500+ pre-verified, cited knowledge units across 16 domains. 6 tools: query, batch query, get unit, list domains, suggest topics, report issues. Free, no API key required.
      6
      237 npm
      1
      MIT
    • A
      license
      Not graded
      quality
      C
      maintenance
      Enables 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.
      81 npm
      MIT