About AIAgentCycle
AIAgentCycle is a read-only MCP server serving current, sourced information about the Dutch company AIAgentCycle. It is for AI agents answering questions about the company's services, integrations, cases, and contact details; website content is synced hourly in Dutch and English.
Key features
- 01Full-text website search
- 02Retrieve page Markdown
- 03List documents by language and type
- 04Look up integrations and status
- 05Query services, prices, and cases
- 06Structured read-only results
Use cases
- 01Answer questions about services and prices
- 02Find supported system integrations
- 03Retrieve published case studies
- 04Look up company and contact details
MCP support
This read-only MCP server provides search, document, integration, service, case, and contact-information tools over Streamable HTTP, returning structured JSON.
Configuration
Remote servers need no local install. Pick your client, then copy the config or install in one click.
https://mcp.aiagentcycle.com/mcpclaude mcp add --transport http com-aiagentcycle-aiagentcycle https://mcp.aiagentcycle.com/mcp{"mcpServers":{"com-aiagentcycle-aiagentcycle":{"type":"http","url":"https://mcp.aiagentcycle.com/mcp"}}}Open in Cursor{"servers":{"com-aiagentcycle-aiagentcycle":{"type":"http","url":"https://mcp.aiagentcycle.com/mcp"}}}Open in VS Code{"mcpServers":{"com-aiagentcycle-aiagentcycle":{"type":"http","url":"https://mcp.aiagentcycle.com/mcp"}}}{"mcpServers":{"com-aiagentcycle-aiagentcycle":{"serverUrl":"https://mcp.aiagentcycle.com/mcp"}}}Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.
Tools
Details
From the project READMEAIAgentCycle MCP server
Read-only Model Context Protocol server for aiagentcycle.com, a Dutch company that builds workflow automation, system integrations and AI agents on self-hosted n8n for small and medium-sized businesses.
AI agents can use it to answer questions about the company with up-to-date, sourced facts: services and prices, the 190 systems it integrates, case studies, the knowledge base and contact details. Content is mirrored from the website every hour, in Dutch (primary) and English.
- Endpoint:
https://mcp.aiagentcycle.com/mcp(Streamable HTTP, no login, read-only) - MCP Registry:
com.aiagentcycle/aiagentcycle - Server Card: https://mcp.aiagentcycle.com/mcp/server-card
- For crawlers and LLMs:
llms.txt,llms-full.txt, Markdown per page at/md/{id}.md
Connect
Add it as a remote MCP server in any client that supports Streamable HTTP, for example in Claude Code:
claude mcp add --transport http aiagentcycle https://mcp.aiagentcycle.com/mcp
Or install the AIAgentCycle plugin from its repository marketplace:
/plugin install aiagentcycle --marketplace aiagentcycle/aiagentcycle-mcp
Tools
| Tool | What it returns |
|---|---|
search_knowledge_base | Ranked full-text search over every page, article, case, solution, sector and integration page |
get_document | Full Markdown text of one page, with source URL and last-modified date |
list_documents | Catalogue of the website by language and type |
list_integrations | The 190 systems AIAgentCycle connects, with status (in production, n8n built-in, custom), domain, category and open-source license |
list_services / get_service | Services and maintenance packages with indicative prices; each price is checked against the live pricing page (price_verified) |
list_cases | Published case studies |
get_faq | FAQ, verbatim from the pricing page |
get_company_profile | Who the company is; key facts checked against the About page (facts_verified) |
get_contact_info | Email, phone and the free assessment page. The server never sends messages or books appointments |
All tools are annotated readOnlyHint: true, return structured JSON, and log only a hash of their arguments.
How it stays correct
sync/aac_mcp_sync.pypulls all public content from the WordPress REST API every hour. It also reads the integration catalogue from the published HTML and the FAQ from the FAQPage JSON-LD. The sync is fail-closed: a partial or broken site never replaces the previous knowledge base.- Hand-maintained facts (prices, profile) carry
verifystrings that must appear verbatim on the live source page. If the site changes, the tools reportprice_verified: falseand a daily check turns red. - A daily check covers reachability, freshness, discovery files, transport rules, TLS and the Registry entry.
Run locally
python -m venv .venv && .venv/bin/pip install --require-hashes -r requirements.txt
AAC_MCP_CONTENT=./content-published python -m aac_mcp --host 127.0.0.1 --port 8321
.venv/bin/python -m pytest -q
For an isolated container build, including directory scanners such as Glama:
docker build -t aiagentcycle-mcp .
docker run --rm -p 127.0.0.1:8321:8321 aiagentcycle-mcp
The regular command only binds to localhost. The container listens on its internal interface but the documented host mapping remains localhost-only. In production the server runs behind nginx with rate limiting and an Origin check. See deploy/ for the systemd units and the nginx config. SCHEMA.md describes the content files.
Built with FastMCP. Maintained by AIAgentCycle (Good Nature Vibe B.V., Oosterwolde, the Netherlands).
License
MIT, see LICENSE. The texts in content-published/ are AIAgentCycle's own business information; they are included as an example of the content format.
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.