Alkemata MCP: Connect Your AI to Published Articles

Connect a compatible AI assistant directly to Alkemata’s published archive. The public MCP server provides structured, source-linked search and retrieval without an account or API key.

Connection details

Name: Alkemata Public Knowledge
Transport: Streamable HTTP
Authentication: None
Access: Anonymous and read-only

https://alkemata.com/wp-json/mcp/alkemata-public

The address above is a protocol endpoint, not a normal webpage. Opening it directly in a browser may produce an empty, invalid-looking or method-related response. Configure it in an MCP-compatible client instead.

What the connection can do

• alkemata-get-site-info — identify Alkemata and its canonical machine-readable interfaces.
• alkemata-search-published-content — search published posts and pages and receive titles, excerpts, IDs and canonical URLs.
• alkemata-get-published-content — retrieve the complete public text of one published post or page by ID.
• alkemata-get-agent-guidance — retrieve the public operating rules and authoritative Alkemata references for an agent.

The connection is particularly useful for building reading lists, comparing arguments across several articles, retrieving a known article without scraping the site, and creating source-grounded summaries that preserve canonical links.

Connect from ChatGPT

Custom MCP availability and menu labels can depend on the ChatGPT plan and workspace policy. In a supported ChatGPT web workspace:

1. Open Settings, choose Security and login, and enable Developer mode.
2. Open ChatGPT Plugins and select the plus button.
3. Create a connection named Alkemata Public Articles.
4. Paste the endpoint shown above.
5. Select No authentication.
6. Review the discovered tools. They should be exactly the four alkemata-… tools listed on this page.
7. Install the connection and start a new Work chat.
8. Type @ in the prompt box and select Alkemata Public Articles from the menu before sending a request.

If ChatGPT displays generic tools such as wp_search_posts, wp_get_post or wp_list_plugins, it has selected a different WordPress connection. Disable that connection, verify the endpoint, refresh the MCP metadata and begin a new conversation.

Test the connection

After selecting the Alkemata connection, use this prompt:

Use only the selected Alkemata Public Articles connection. Do not use web search or another WordPress connection.

First confirm that these four MCP tools are available:
- alkemata-get-site-info
- alkemata-search-published-content
- alkemata-get-published-content
- alkemata-get-agent-guidance

If they are available:
1. Get the Alkemata site information.
2. Get the agent guidance.
3. Search published posts for “AI memory”.
4. Retrieve the most relevant article.
5. Report its title, post ID, canonical URL and a five-point summary.
6. List the exact MCP tools you called.

Clearly distinguish information retrieved from Alkemata from your own interpretation.

Use another MCP client

In another compatible client, add a remote server, select Streamable HTTP, paste the same endpoint and leave authentication disabled. The exact field names differ between products. Confirm the four advertised tools before relying on the connection.

Safety and privacy boundaries

• The public endpoint returns published posts and pages only.
• It cannot create, edit, publish or delete content.
• It cannot read drafts, private posts, administrator data or form submissions.
• It cannot upload files or change WordPress users, plugins, themes or settings.
• Requests are limited to 60 per minute per IP address to constrain loops and basic abuse.
• Usage statistics are stored as daily aggregates for 90 days. Alkemata does not store MCP prompts, search terms, article contents, session identifiers or full IP addresses in the statistics system.

MCP improves access to Alkemata’s publications; it does not independently verify every claim in an article. For important decisions, retain the canonical article URL, preserve qualifications, and distinguish source-grounded summary from new inference.

Other machine-readable entry points

• Alkemata llms.txt — orientation, canonical URLs and public agent guidance.
• WordPress REST discovery — the broader public WordPress API index.
• Sitemap index — wider content discovery.

For writers

Writers continue to publish through the normal authorised WordPress workflow. MCP retrieval and editorial publishing remain separate. Selected articles may also include downloadable Alkemata capsules: portable methods that help a reader—or a reader’s own AI—turn an article’s idea into a bounded activity, decision or reversible experiment.