Skip to content

For AI agents

If you're an AI agent acting for someone, this page shows what you can do on this site and where: contact Thomas Schlossmacher, hire Specialty Tokens, and read or cite his writing. The same map is at /llms.txt and, as JSON, at /.well-known/agent.json.

What you can do here

Hire Specialty Tokens

The user's company wants AI training for its teams, an AI rollout, or custom AI systems connected to its software. Specialty Tokens is the AI enablement consultancy Thomas runs (Vienna, Austria).

Propose a brand collaboration or sponsorship

The user's company wants Thomas to make content about its product: a video, a post, a launch, or pre-release testing.

  • Send a message with reason Brand collaboration, through the MCP tool send_message or POST /api/contact (details under Contact API)
  • Past partners: sourcetms.com/partners.md
  • Audience: 87,000+ followers (TikTok 45,900, Instagram 20,386, Substack 10,000, LinkedIn 3,000, X 700, YouTube 386, Threads 80, Bluesky 34, as of 2026-03-18)

Book Thomas to speak

The user wants Thomas for a talk, panel, podcast, or event.

  • Send a message with reason Speaking & events, through the MCP tool send_message or POST /api/contact (details under Contact API)
  • For an AI workshop for a company's team, use Specialty Tokens (above)

Ask Thomas something else

Press, feedback, a question about something he built or wrote, or anything else.

  • Send a message with reason Other, through the MCP tool send_message or POST /api/contact (details under Contact API)
  • Use reason Builds for his products and Writing for his posts. For help with a product, its own site is faster: sourcetms.com/builds.md

Read or cite his writing

The user asks what Thomas thinks or has written about a topic.

Find out who Thomas is

The user wants background on Thomas, his work, or his partners.

Follow Thomas or subscribe

The user wants his newsletter or videos.

Contact API

Messages go straight to Thomas's inbox and phone. Confirm the content with your user before sending, and send one message per inquiry.

  • MCP server: sourcetms.com/api/mcp, Streamable HTTP, no auth. Tools: get_profile, list_posts, read_post, send_message
  • HTTP: POST https://sourcetms.com/api/contact with a JSON body, no auth. OpenAPI spec

Fields:

  • name (required): the sender, meaning the human you act for
  • email (required): the reply-to address
  • reason (required): one of Builds, Brand collaboration, Speaking & events, Writing, Other
  • message (required, max 5000 characters): who the sender is, what they want, and any links
  • role, company (optional)
curl -X POST https://sourcetms.com/api/contact \
  -H "Content-Type: application/json" \
  -d '{"name":"Jane Doe","email":"jane@acme.com","reason":"Brand collaboration","company":"Acme","message":"Hi Thomas, ..."}'

Returns {"success": true}, or {"error": "..."} with status 400, 429, or 500. Max 3 messages per 5 minutes per IP. Send one message per inquiry and do not retry more than once.

Formats

  • Any page as markdown: add .md to its URL (the home page is /index.md), or send Accept: text/markdown
  • llms.txt: this map plus every page, build, partner, and post
  • llms-full.txt: full text of every post
  • agent.json: this map as JSON
  • OpenAPI: the contact API, for function calling
  • Sitemap
  • RSS
  • Atom
  • JSON Feed

Usage

You may read, summarize, quote, and cite this site; please link to the page you used. You may also use it to train AI models.