Peptide Library MCP server

Bring Peptide Library's research profiles into the AI assistant you already use. The server answers with structured, cited data from the Peptide Library and links every result back to its page.

Server URL

https://peptidelibrary.app/mcp

Streamable HTTP · no authentication · read-only · version 1.0.1

What it does

Ten read-only tools, all working on Peptide Library's own reviewed data:

  • search_peptides Search peptides. Fuzzy search by name, synonym or alias, with research-category and research-area filters and pagination.
  • get_peptide_details Peptide research profile. Classification, mechanism and targets, evidence by tier (regulatory label, controlled human trials, other human studies, animal and in-vitro research) with cited sources, safety, FDA/WADA status and review dates.
  • compare_peptides Compare peptides. Side-by-side comparison of 2 or 3 profiles on structured fields, marking what is the same, different or undocumented. Never ranks.
  • search_research Search educational content. Search Peptide Library's FAQ, per-peptide FAQs, articles and courses, with links and attribution.
  • get_peptide_categories Research categories. The library research categories with descriptions, counts and the profiles in each.
  • get_research_trends Research trends. This week's most-researched peptides in the community, as privacy-preserving percentages (no head counts).
  • convert_units Convert units. mg ↔ mcg exactly, and IU only with a peptide's documented IU-per-mg factor.
  • solution_concentration Solution concentration. Amount ÷ volume as mg/mL, mcg/mL (and IU/mL with a documented factor).
  • blend_concentrations Blend concentrations. Each component's concentration and share when several are dissolved in one volume.
  • cost_per_amount Cost per amount. Total cost and cost per mg, mcg or IU from a price you enter. No price estimates or currency conversion.

It also offers resources (peptidelibrary://categories, peptidelibrary://evidence-levels, peptidelibrary://methodology, peptidelibrary://glossary, peptidelibrary://disclaimer, peptidelibrary://peptide/{id}) and prompts that keep answers grounded: compare peptide research, summarize peptide evidence, explain a mechanism, find research topics, understand limitations.

Questions to try

  • What does the research say about BPC-157, and how much of it is in humans?
  • Compare semaglutide and tirzepatide: mechanism, evidence and regulatory status.
  • Which peptides in the library are studied for cognition?
  • What is the FDA and WADA status of bremelanotide (PT-141)?
  • Explain how GLP-1 receptor agonists work, citing Peptide Library.
  • If 10 mg is dissolved in 2 mL, what is the concentration in mcg/mL?
  • Which peptides are researchers exploring most this week?

Connect your assistant

Claude (web and desktop)

Open Customize → Connectors, choose Add custom connector, name it “Peptide Library” and paste the server URL. On Team and Enterprise plans an owner adds it under Organization settings → Connectors. Free plans allow one custom connector.

Claude Code

claude mcp add --transport http peptide-library https://peptidelibrary.app/mcp

ChatGPT

Where your plan or workspace allows custom MCP servers: open Plugins, choose Add custom MCP server, paste the server URL, select No authentication and create it. Then type @ in a chat to use it. Menu names change between ChatGPT releases; see OpenAI's help center if yours differ.

Cursor

Add to ~/.cursor/mcp.json (or .cursor/mcp.json in a project):

{
  "mcpServers": {
    "peptide-library": {
      "url": "https://peptidelibrary.app/mcp"
    }
  }
}

VS Code

Add to .vscode/mcp.json:

{
  "servers": {
    "peptide-library": {
      "type": "http",
      "url": "https://peptidelibrary.app/mcp"
    }
  }
}

Other clients and the MCP Inspector

Any client that supports remote Streamable HTTP servers can use the URL directly. To explore the tools:

npx @modelcontextprotocol/inspector --cli https://peptidelibrary.app/mcp --transport http --method tools/list

Endpoint details

  • Transport: Streamable HTTP at https://peptidelibrary.app/mcp (POST). Stateless: protocol revision 2026-07-28 and 2025-era clients (2024-11-05 to 2025-11-25) are both served.
  • Authentication: none. Every tool is read-only and works anonymously.
  • Limits: about 300 requests per minute per IP address, 64 KB per request, 25 results per page.
  • Results carry structured JSON (with an output schema per tool), the Peptide Library URL of the source page and a research-use disclaimer. Fields the library lacks are returned as not_documented.

What it will not do

  • Give doses, regimens, cycles, reconstitution or injection instructions.
  • List vendors, prices or purchase links, or help anyone buy anything.
  • Read or change any account, tracker, inventory or other personal data.
  • Rank peptides, or invent studies or citations: missing data stays missing.

Research and educational use only. Not medical advice. Peptide Library does not sell peptides.

Privacy and security

The server needs no account and stores no personal data. Peptide Library keeps daily counts of requests, tool calls, errors and the client name your assistant reports, never tool arguments, results or IP addresses. Our host, Netlify, processes IP addresses to apply rate limits. Your AI provider's own privacy terms cover what you type into it. See the privacy policy.

Report a security issue to support@peptidelibrary.app.

Where the data comes from

Profiles are written and reviewed under Peptide Library's editorial policy and methodology: claims cite PubMed, DOI, trial registries and FDA or WADA records, and evidence is labelled by kind (human trials, animal and lab research). Trends come from the Peptide Trends rollup and show percentages only.

Use Peptide Library directly

The website and the iOS and Android apps add calculators, a research tracker, comparisons and the full Peptide Trends dashboards.

FAQ

What is the Peptide Library MCP server?

A public Model Context Protocol endpoint that lets AI assistants look up Peptide Library research profiles, compare peptides, search its educational content and do unit arithmetic, with links back to the source pages.

Do I need a Peptide Library account?

No. The server is free, anonymous and read-only. It never reads accounts, trackers, inventory or any personal data.

Does it give dosing advice?

No. It leaves out every dose figure, regimen, reconstitution and injection instruction, and it does not list vendors or prices. It is research education, not medical advice.

Which AI clients work with it?

Any client that supports remote MCP servers over Streamable HTTP, including Claude (web, desktop and Claude Code), ChatGPT where custom MCP servers are enabled for your plan, Cursor and VS Code.

Where does the data come from?

The same database as peptidelibrary.app and the Peptide Library apps: editor-reviewed profiles that cite PubMed, DOI, trial-registry and regulatory sources. Each result links to its page.

What is logged when I use it?

Daily counts of requests, tool calls and errors and the name your AI client reports (for example "claude-ai"). Tool arguments, results and IP addresses are not stored by Peptide Library; Netlify sees IP addresses to enforce rate limits.

Client configuration examples and registry metadata: github.com/ryandelauz/peptide-library-mcp. Questions or problems: support or support@peptidelibrary.app.