Agent Readable Devcon --> an open knowledge base so attendees' AI agents can navigate Devcon

Devcon Improvement Proposals (DIPs)
1

Title: Agent-Readable Devcon — an open, machine-readable knowledge base so attendees’ AI agents can navigate (and plan for) Devcon

Authors: Anand Vashishtha — https://x.com/AnandVashisht15

Resources Required: Mostly digital. I need read access to the official schedule / venue / FAQ / travel data (or permission to crawl the public site), and the OK to surface a link or QR code in the Devcon app or on community-hub signage. Optional, for a live “ask-an-agent” demo: a small table, power, and one screen. No dedicated operations support needed.

:brain: What do you suggest?
Publish Devcon’s schedule, venue map, FAQs and travel basics as an open, machine-readable knowledge base — a single agent-friendly endpoint (plus an /llms.txt index and structured data) that any attendee’s AI assistant can read. Scan a QR or open a link, then just ask your own agent: “what’s on now, where is it, how do I get there, and what should I book?” No digging through the app.

:bullseye: Motivation & Rationale

  • Agents are becoming our coworkers. In 2026, people increasingly hand planning, scheduling and booking to an AI assistant. For an event the size of Devcon, attendees will point their agent at it — for trip planning, ticket choices, and on-the-day navigation — but only if the data is in a form an agent can actually read. Today most event info lives in human-only UIs and JavaScript apps that agents parse poorly, so the agent guesses or gives up.
  • This is “Agentic Engine Optimization” for an event. Just as we once learned SEO for search crawlers, sites now have to be readable by autonomous agents: clean structured data (schema.org), an /llms.txt index, stable endpoints, high text-to-markup ratio. Devcon adopting this makes it the reference example of an agent-ready Ethereum event.
  • Better experience for everyone. Devcon is huge and multi-track; finding the right talk or room fast is hard. Asking your own assistant — including in regional Indian languages — is faster and more accessible than scrolling an app, and it especially helps first-timers and non-English speakers.
  • On-thesis for where Ethereum is going. Ethereum is actively building the agent economy: ERC-8004 “Trustless Agents” (on-chain agent identity + reputation, now live on mainnet), A2A and MCP for agent interoperability, and agentic payment rails like x402. An open, agent-readable Devcon is a concrete, hands-on showcase of exactly that direction — and the dataset + tooling are open public goods anyone can fork and reuse at future Ethereum events, not a closed app feature.

:hammer_and_wrench: Implementation

  • Done before at an event? Not yet — this would be the first run. But the core already exists: I run 0gskills (https://0gskills.com), an open agent-readable knowledge base, so I’m adapting working tooling rather than starting from zero. (Background: I build open agent infrastructure and won an x402 agent-payments hackathon challenge.)
  • How I’ll make it happen:
    1. Ingest the official schedule / venue / FAQ / travel data into clean, structured, machine-readable formats (JSON + Markdown), following emerging agent-readability conventions: schema.org structured data, an /llms.txt index, and a high text-to-markup ratio.
    2. Expose it behind a simple, stable, open read endpoint any agent can hit — plus an optional MCP-style interface so MCP-capable assistants can query it directly (“what AI talks are next?”, “fastest route to Hall 2?”).
    3. Keep it auto-synced as the schedule changes during the event.
    4. Surface access via a QR / link in the Devcon app or on community-hub signage (model: the accepted Agora DIP, which only needed QR placement).
    5. Optional: a small “ask-an-agent” demo corner + a short guide on making your own project agent-readable.
    6. Post-event: publish the full dataset + tooling openly under a permissive license, with a short writeup and lightweight usage metrics (scan counts, sample queries).

:gear: Operational Requirements & Ownership

  • What’s needed at Devcon: read access to (or permission to crawl) the official event data, and the OK to display a link / QR. Everything else is digital and hosted by me — no physical build, no staffing burden on the Devcon team.
  • Who runs it on site: me — Anand Vashishtha (contact below). Solo builder and maintainer, available all event days to keep it synced and fix issues.
  • Privacy & scope: read-only public event data only — no attendee personal data, no payments handled by the KB itself. Fully open-source so anyone can audit it.
  • Could integrate with: the Devcon app / data team (data source + link surface), AI and public-goods community hubs, and any on-chain-agent or agent-tooling project (e.g. the ERC-8004 / MCP / A2A ecosystems) that wants its own docs made agent-readable.

:link: Links & Additional Information

2

This is an interesting proposal, especially as more people start relying on AI agents for planning and navigation. Making Devcon data machine-readable through standards such as schema.org and llms.txt could significantly improve how agents discover schedules, venues, and event information. The idea also aligns with the broader trend of AI-powered knowledge systems and agent-based interactions, which are becoming increasingly important as AI capabilities evolve. For anyone interested in the underlying concepts, this overview of Machine Learning provides some useful background on how intelligent systems process and learn from data. Overall, an open and reusable approach like this could become a valuable reference for future conferences and community events.