devoxx-voxxed-cfp

Devoxx / VoxxedDays — CFP API

Answer the user's question by querying public, no-auth REST APIs. Each Devoxx and VoxxedDays event has its own CFP instance at https://{event-slug}.cfp.dev/api/public. Fetch live data with curl in the bash tool and parse JSON with python3. Never invent talk titles, times, rooms, or speaker facts — always pull from the API.

Step 0 — resolve the event slug

Discover events from the Devoxxians registry (no auth):

• GET https://devoxxians.com/api/public/events/upcoming
• GET https://devoxxians.com/api/public/events/past

Each event has name (e.g. "Devoxx UK 2026"), eventCategory (DEVOXX or VOXXED), fromDate/toDate, location, website, and apiURL (e.g. https://devoxxuk26.cfp.dev/api/). The slug is the hostname prefix of apiURL — extract with re.match(r"https://([^.]+)\.cfp\.dev", apiURL).

Resolution rules:

• Match the user's event mention against name case-insensitively (fetch both lists; "Devoxx UK" without a year → prefer the nearest upcoming edition, else the most recent past one).
• No event mentioned at all → use conversation context; if still ambiguous, ask which event (and offer the upcoming list).
• apiURL can be null for announced events whose CFP isn't live yet (e.g. a next-year edition) — say so and fall back to the most recent edition that has an apiURL.
• Cache the slug for the rest of the conversation; don't re-resolve on every question.

Questions like "when is the next Devoxx" or "which Voxxed events are coming up" are answered directly from these two registry endpoints.

Per-event endpoints

Base: https://{slug}.cfp.dev/api/public — all GET, no auth.

Need	Endpoint
Event metadata + CFP open/close dates	/event
All talks (content)	/talks
All speakers (summary, paginated)	/speakers?page=N
One speaker + their talks	/speakers/{id}
Available schedule days	/schedules (returns day links — days VARY per event, never assume monday–friday)
Schedule for a day	/schedules/{day}
Rooms / Tracks	/rooms, /tracks

Full field reference, gotchas, and worked curl+python3 recipes are in references/api.md — read it before composing a non-trivial query.

Key joins

• A schedule slot's proposal.id equals a talk's id in /talks.
• A talk's speakers[] and a speaker's proposals[] link talks ↔ speakers.
• /talks carries content (description, track, level, favourites) but its timeSlots is usually empty — get times and rooms from /schedules/{day}.

Times

Display times in the event's local timezone. Every schedule slot carries fromDate/toDate in UTC plus a timezone field (e.g. Europe/London, Europe/Brussels). Convert with zoneinfo:

from datetime import datetime
from zoneinfo import ZoneInfo
def local(ts, tz): return datetime.fromisoformat(ts.replace("Z","+00:00")).astimezone(ZoneInfo(tz)).strftime("%H:%M")

Linking to the public website (mandatory)

Every talk and every speaker referenced in an answer MUST be a clickable markdown link to https://m.devoxx.com, using the same event slug:

• Talk: https://m.devoxx.com/events/{event-slug}/talks/{talkId}/{slug(title)} — e.g. Top 10 Event-Driven Architecture Pitfalls
• Speaker: https://m.devoxx.com/events/{event-slug}/speaker/{speakerId}/{slug(fullName)} (note: speaker, singular) — e.g. Victor Rentea

Routing is by id (the trailing slug is cosmetic), but generate it properly:

import re
def slug(s): return re.sub(r"[^a-z0-9]+", "-", (s or "").lower()).strip("-")

Format as [Talk Title](url) and [Speaker Name](url) everywhere — including lists and schedule overviews. For event-level answers, link the event's website from the registry payload.

Answering patterns

• "Tell me about " → page through /speakers to find the id, then GET /speakers/{id} for bio + proposals.
• "Who's speaking about " / "find talks" → fetch /talks, case-insensitively match across title, summary, description, keywords[].name, track.name.
• "What's on " / "what's in " → GET /schedules for valid days, then /schedules/{day}, filter by room, sort by fromDate.
• "Most popular talks" → sort /talks by totalFavourites descending.
• "When/where is the next Devoxx or Voxxed event" → registry endpoints on devoxxians.com.
• "When does/did the CFP open or close?" / "is the CFP open for ?" → resolve the slug, then GET /event and read cfpOpening/cfpClosing (UTC ISO). Compare to now to label open / closed / opens-later; an event whose apiURL is null has no CFP instance yet. The registry does NOT carry CFP dates — always use /event.

Keep answers focused on what was asked. Every talk title and speaker name in the answer must be a markdown link per the linking rules above.