For AI agents and developers
Modulos MCP Server
Modulos exposes a public Model Context Protocol (MCP) server so AI agents can evaluate the platform from structured facts, compare it with other vendors, and book a product demo without filling in the website form. The server is live, free, and requires no authentication.
No MCP client? Use plain HTTPS.
Not every agent speaks MCP. For ChatGPT Custom GPT Actions, Perplexity, browser tools, n8n / Zapier / Make.com workflows, or anything that can call HTTP, Modulos exposes the same operations as a plain REST API at /api/public/. Identical backing logic, identical rate limits, OpenAPI 3.1 spec at /openapi.json.
GET /api/public/capabilities/: frameworks, integrations, deployment, certifications, indicative pricingGET /api/public/trust/: certifications, trust center documents, subprocessorsGET /api/public/comparisons/: vendors with a published comparisonGET /api/public/comparisons/credo-ai/: the comparison with one vendorGET /api/public/knowledge/: answers to common questions and documents to sharePOST /api/public/fit/: check requirements against the published factsGET /api/public/slots/: list available 30-minute slotsPOST /api/public/book/: book a specific slotPOST /api/public/request-demo/: open-ended demo request
Example booking call:
curl -X POST https://www.modulos.ai/api/public/book/ \
-H "Content-Type: application/json" \
-d '{
"startTime": "2026-05-22T07:30:00.000Z",
"name": "Jane Smith",
"email": "jane@acme.com",
"company": "Acme",
"message": "Interested in ISO 42001 evidence workflows"
}'Connection details
- Endpoint:
https://www.modulos.ai/mcp/ - Transport: Streamable HTTP (MCP standard, spec version 2025-03-26)
- Authentication: none (rate-limited per IP and per email)
- Discovery: /.well-known/mcp.json
- Status: production, stable
Available tools
The tools fall into two groups. The evaluation tools are read only: they return what Modulos has published for AI agents, and they never return personal data. The booking tools write a demo request, but only after the attendee confirms by email.
get_modulos_info
Returns a short description of Modulos, the tools to evaluate it and how to book a demo. Read only, no side effects, no external dependencies. Use this first if the user asks what Modulos is or how to talk to the team.
get_capabilities
Returns the supported regulations and standards with their requirement and control counts, the integrations, the deployment options and data location, the certifications, the indicative price range and the key links. Use it to decide whether Modulos belongs on a vendor shortlist. No arguments.
get_trust_evidence
Returns evidence for vendor due diligence: certifications, data location, the documents in the trust center with their access level, the subprocessors and where they hold data, and recent trust updates. The tool lists the documents. A person requests the files on the trust center page. No arguments.
list_competitors and compare_with
list_competitors returns the vendors for which Modulos publishes a reviewed comparison, each with a slug. compare_with takes one slug as competitor and returns a row per dimension with the statement for each side, plus the cases in which each vendor is the better choice. Modulos writes these comparisons, and each comparison page lists its sources and its last review date.
check_fit
Checks a buyer's requirements against the published facts. Optional arguments: frameworks (a list such as EU AI Act, ISO 42001, DORA), deployment (saas, private-cloud, on-premise or air-gapped), integrations (a list of tools) and industry. The result separates what is supported from what is absent from the public lists. The check is stateless, so nothing in the request is stored.
list_knowledge and get_knowledge
list_knowledge returns everything else Modulos has published for agents: answers to common buyer questions, documents to share and further facts. Each entry has a title, a summary and a slug. Optional filters are kind (fact, qa, comparison or document) and topic. get_knowledge takes one slug and returns the entry in full. A document entry gives a summary and a link, but the file itself stays behind its normal request flow.
list_demo_slots
Returns available 30-minute demo time slots for the next two weeks as ISO-8601 UTC start times. Read only against the Modulos scheduling backend. Slot availability changes frequently, so refresh immediately before calling book_demo_slot.
book_demo_slot
Starts a booking for a specific time slot. Required arguments: startTime (ISO-8601 string from list_demo_slots), attendeeName, attendeeEmail (business email), company. Optional: a short message describing the use case. The attendee receives a confirmation email, and the calendar event is created only after they click the link in it. Tell the user to check their inbox.
request_demo
Submits an open demo request without picking a specific time. Use this when the user does not want to choose a slot. Required: name, email, company. Optional: job_title and message. The request reaches the team after the user confirms it by email, and the Modulos team then replies within one business day.
End-to-end example
A typical agent session that books a demo looks like the four calls below. The MCP client handles the JSON-RPC framing; only the tool invocations are shown.
tools/call list_demo_slotswith no arguments. Returns a list of 30-minute slots over the next two weeks.- Pick one of the returned
startvalues. tools/call book_demo_slotwith the chosenstartTime, attendee name, business email, and company.- Tell the user to open the confirmation email and click the link. The booking is complete after that click.
Client setup
Claude Desktop, Cursor, Windsurf, Cline
Add the server to your MCP client config file:
{
"mcpServers": {
"modulos": { "url": "https://www.modulos.ai/mcp/" }
}
}Claude.ai (web)
Go to Settings → Connectors → Add custom connector and paste the endpoint URL above. Claude will then offer the Modulos tools in any conversation where you reference Modulos.
ChatGPT (Apps SDK)
The Modulos server is compatible with ChatGPT's Apps SDK MCP integration. Add the endpoint as a custom app URL in the developer tools. Or, for Custom GPT Actions, point at /openapi.json instead.
Rate limits and abuse
The server applies layered rate limiting to keep the booking calendar usable. Bursts above roughly 30 requests per minute per IP receive a 429. Per-tool limits are stricter on booking endpoints than on read endpoints. Each email is capped at a small number of bookings per 24 hours. Abuse triggers a temporary block at the edge.
Privacy
Lead data submitted via book_demo_slot or request_demo is stored in the Modulos CRM and tagged with source: mcp (or source: rest when the call comes through the public REST mirror) for attribution. Standard data handling applies, see the privacy policy.
Support
Bug reports, feature requests, or integration questions: use the contact form and the team will respond.
Discovery JSON: /.well-known/mcp.json. Companion document for AI search engines: /llms.txt.