Skip to content
MCP server

Catalyst evidence graph

Read-only MCP server returning sourced findings on what supplements and compounds do

Pricing
Free
Platforms
Web, API, CLI
Runs
Remotely
License
MIT
Version
0.2.0
Listing updated
3 days ago

About Catalyst evidence graph

Catalyst evidence graph is a free MCP server for health and science. It runs as a hosted remote server, so there is nothing to install locally.

Each finding comes with an evidence strength of strong, moderate, limited, very limited or insufficient. That level is computed from the study behind the finding rather than set by hand, and the response also includes the verbatim sentence it came from, the paper, the population and dose, and a permalink for checking. Relations from reference databases, such as targets, pathways and reactions, are labelled as hypotheses rather than results.

The project states that evidence strength is not a recommendation and not medical advice. No tool returns a product, a price or a buy link, and none ranks what to take. The server stores no queries and has no accounts. Version 0.2.0 replaced letter grades with named levels, and get_findings accepts an evidence_strength floor.

The repository, licensed MIT, holds client setup for Claude, ChatGPT, Claude Code, VS Code with Copilot, Cursor and Gemini CLI. The graph content itself follows separate terms.

Install

Run in a terminal

claude mcp add --transport http evidence-graph https://catalystproject.ai/mcp

Add to Settings → Connectors → Add custom connector

https://catalystproject.ai/mcp
Install in CursorOpens Cursor, which asks you to confirm. Or add it by hand:

Add to ~/.cursor/mcp.json

{
  "mcpServers": {
    "evidence-graph": {
      "url": "https://catalystproject.ai/mcp"
    }
  }
}
Install in VS CodeOpens VS Code, which asks you to confirm. Or add it by hand:

Add to .vscode/mcp.json

{
  "servers": {
    "evidence-graph": {
      "type": "http",
      "url": "https://catalystproject.ai/mcp"
    }
  }
}

Add to ~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "evidence-graph": {
      "serverUrl": "https://catalystproject.ai/mcp"
    }
  }
}

Add to your client's MCP settings (most use this shape)

{
  "mcpServers": {
    "evidence-graph": {
      "type": "http",
      "url": "https://catalystproject.ai/mcp"
    }
  }
}

Tools this server provides (7)

Read directly from the server.

find_evidence
Answers "what does X do?" and "does X affect Y?" for a supplement, nutrient, drug, food compound or plant (for example "does magnesium help sleep?", "what is creatine shown to do?", "omega-3 and triglycerides"). Pass the substance as query and the effect or outcome asked about, if there is one, as about. When the person names two substances ("is creatine safe with caffeine"), pass one as query and
search_nodes
Looks up the node id for a compound, protein, organism or outcome by name or synonym (for example "CoQ10", "caffeine", "CYP1A2", "sleep"). Use it to get the id that get_findings, get_finding, get_relations, get_reactions and reach take; for a question about what a substance does, call find_evidence instead, which needs no id. Returns up to 10 hits, each with id, label, type and the number of findi
get_findings
Lists the curated study findings for one node by id, from search_nodes: for a compound, organism or material, what it has been shown to affect; for an outcome, which compounds have been studied against it. Use it to browse a node by id. For a question in words, such as "does X help Y", find_evidence answers in one call and needs no id. summary has one row per thing the node has findings about, cou
get_finding
Explains one finding by its finding id: the claim, its evidence strength and what that level means, the verbatim quote, the study design and the paper, and evidence_strength_reasons, each rule that set the level in order, and what the assessment does not weigh. The finding id is 26 characters, from the id field of a find_evidence or get_findings finding, or the last part of its permalink (/e/{id})
get_relations
Lists what public reference databases (Reactome, UniProt, Rhea, ChEBI and others) record about one node by id, from search_nodes: transporters, enzymes, pathways, expression. Every row is assertion_class 'inferred': a hypothesis about how a compound could act, not a result anyone measured in people. Present these as possible mechanisms and never as effects. For effects, use find_evidence. For the
get_reactions
Lists the biochemical reactions from Rhea and Human-GEM that a compound is a substrate or product of, or that a protein (enzyme) catalyses or transports, by node id from search_nodes. It lists reactions and does not answer a route question: for "can X become Y" or "what can X become", use reach. Every row is assertion_class 'inferred': a reaction two public databases record, not a result anyone me
reach
Answers a route question: "can X become Y in the body" (give from_id and to_id) and "what can X become" (give only from_id). Both ids are compound node ids from search_nodes, so a question that names two compounds needs one search_nodes call for each first. It searches the reactions two public databases (Rhea, Human-GEM) record, up to depth steps (1 to 3, default 3), and returns the routes with th

What it does

Evidence strength levels
Five named levels computed from the underlying study.
Verbatim source quotes
Each finding carries the original sentence and the paper.
Population and dose
Findings state who was studied and at what dose.
Hypothesis labels
Database relations are marked as hypotheses, not results.
No commercial output
No tool returns products, prices or buy links.
Many client installs
Ships a Claude Code plugin, a VS Code extension and a Gemini CLI extension.

Good for

  • Checking what research says a supplement does before discussing it
  • Pulling cited findings for a nutrient or food outcome
  • Filtering to findings above a minimum evidence strength
  • Giving an assistant sourced evidence instead of unsourced claims

Pricing

Free

Hosted server needs no account or key

Prices change. Check catalystproject.ai before you buy.

Catalyst evidence graph FAQ

Does it need an account or key?

No, the server is read-only with no account and no key.

Is the output medical advice?

No, the project says evidence strength is not a recommendation.

Does it store my queries?

No, the README says it stores no queries.

What licence is the code under?

MIT for the repository code; graph content has its own terms.

Is Catalyst evidence graph free?

Yes. Catalyst evidence graph is free to use.

How do I add Catalyst evidence graph to Claude Code?

Run this command in your terminal: claude mcp add --transport http evidence-graph https://catalystproject.ai/mcp

What are the best alternatives to Catalyst evidence graph?

The closest alternatives to Catalyst evidence graph are Tooluniverse, Oya Browser, Later Stages, MedXpert 标准检索连接器 and MedXpert 全球法规连接器 (reg-connector).

Alternatives to Catalyst evidence graph

  1. Tooluniverse

    MCP server

    MCP server giving AI agents access to thousands of scientific research tools

    Runs locally
    • Open source
    • 1.8K GitHub stars
    • Health and science
  2. Oya Browser

    MCP server

    HIPAA-oriented agentic browser with replayable playbooks, reachable over MCP

    • 32 tools
    • Remote
    • Freemium
    • 349 GitHub stars
    • Browser automation
  3. Later Stages

    MCP server

    California senior living and care provider records from state licensing and Medicare

    • 6 tools
    • Remote
    • Free
    • Health and science
  4. Read-only local index of medical-device standards with official links, no standard text

    Runs locally
    • 4 tools
    • Free
    • Legal and government
  5. Splyt

    MCP server

    Let Claude or ChatGPT read your Splyt workouts, records, plan, habits and sleep

    • Remote
    • Free
    • Health and science