For AI agents
Built to be read by machines too
Everything on this site is available to AI assistants in clean formats: structured data, Markdown, calendar feeds and an MCP server.
Start here
- /llms.txt
A map of the site written for language models
- /llms-full.txt
The whole programme in one plain-text file
- /index.md
The homepage as Markdown
JSON feeds
- /api/editions.json
Both editions: dates, venues, audiences
- /api/sessions.json
267 sessions with speakers, stages and times
- /api/speakers.json
426 speakers with bios
- /api/tracks.json
Themes and tracks per city
- /api/passes.json
Passes and what they include
- /api/partners.json
Partners by tier
Calendar
- /agenda.ics
Every session, subscribable
- /agenda/cairo.ics
Cairo only
- /agenda/alexandria.ics
Alexandria only
- /agenda/opening-remarks-from-heritage-to-innovation-building-the-middle-east-s.ics
One session (every session has one)
Markdown
- /speakers/h-e-mr-ahmed-kouchouk.md
Every speaker page has a .md twin
- /agenda/opening-remarks-from-heritage-to-innovation-building-the-middle-east-s.md
So does every session
Model Context Protocol
A public MCP server
Add this address to Claude, ChatGPT or any MCP client and the assistant can search the agenda and speakers directly. Read-only, no sign-in.
/mcp list_sessions(city, day, track, query) Search and filter the agenda get_session(slug) Full details of one session list_speakers(query) Search speakers by name, company or country get_speaker(slug) A speaker’s bio and their sessions list_passes() Passes and inclusions get_event_info(city) Dates, venue and overview for an edition And on every page
- schema.org structured data (Event, Person, Offer) embedded in the page.
- All content is in the HTML itself. Nothing needs JavaScript to appear.
- Stable URLs for every session and speaker, and agenda filters live in the URL.
- Forms with real labels and autocomplete hints, so an assistant can complete a booking.
- English and Arabic at separate URLs, linked with hreflang.