Skip to main content
Glama

apple-event-mcp

MCP server and webhook worker for Apple event liveblog updates.

It pulls Apple event coverage from public liveblog and news pages and exposes structured posts to AI agents. It can also poll continuously, match posts against user interests, and send signed webhooks.

Sources

  • Engadget: schema.org LiveBlogPosting, structured posts, per-post images

  • MacRumors: static live coverage, timestamped posts, keynote images

  • iClarified: dense transcript-style keynote feed

  • Macworld: commentary-style liveblog entries

  • 9to5Mac: timestamped keynote screenshots, including multi-image feature overviews

  • The Verge: Apple event coverage

  • Any other public HTTPS liveblog or news page supplied through a tool's url argument

The built-in feeds currently target Apple's September 9, 2026 “Surprise and shine” event. The generic URL reader prefers schema.org LiveBlogPosting data and falls back to article markup, so agents can also pass pages from publishers such as The Verge or MacRumors without a dedicated adapter. URLs must use HTTPS, resolve to public IP addresses, and cannot contain credentials or custom ports.

Related MCP server: mcp-ai-news-server

Install

npm install
npm run build

MCP

Run the stdio MCP server:

npm run dev:mcp

Built command:

npm run build
node dist/mcp/server.js

Example MCP client config:

{
  "mcpServers": {
    "apple-event": {
      "command": "node",
      "args": ["/absolute/path/to/apple-event-mcp/dist/mcp/server.js"]
    }
  }
}

Remote HTTP server

The remote server uses the MCP Streamable HTTP transport and requires a Bearer token. Generate a token, then run it locally:

export APPLE_EVENT_MCP_TOKEN="$(openssl rand -hex 32)"
npm run dev:http

Environment variables:

  • APPLE_EVENT_MCP_TOKEN (required): Bearer token accepted by /mcp

  • APPLE_EVENT_MCP_HOST: bind address; defaults to 127.0.0.1

  • APPLE_EVENT_MCP_PORT: listen port; defaults to 8080

  • APPLE_EVENT_ALLOWED_HOSTS: comma-separated HTTP Host allowlist

Connect an MCP client to https://your-host.example/mcp with this header:

Authorization: Bearer YOUR_TOKEN

The unauthenticated GET /healthz endpoint only returns service health. All MCP requests go to POST /mcp and require the token.

Docker deployment

Copy .env.example to .env, replace the token, then run:

docker compose up -d --build

The included Compose file binds the container to 127.0.0.1:8084, ready for a TLS reverse proxy or tunnel.

The production deployment on DevPi is available at:

https://apple-events.shmob.xyz/mcp

Tools:

  • list_events

  • list_liveblog_sources

  • get_liveblog_posts

  • search_liveblog_posts

  • get_event_summary

  • get_noteworthy_updates

Resources:

  • apple-event://september2026/live

  • apple-event://september2026/summary

Every post/search/summary tool accepts either a built-in source or a public HTTPS url. When url is provided, it overrides source:

{
  "url": "https://www.theverge.com/apple-event",
  "limit": 50
}

Webhook Worker

Copy and edit the config:

cp config.example.json config.json

Run once:

npm run dev:worker -- --config config.json --once

Run continuously:

npm run dev:worker -- --config config.json

Replay Mode

Replay mode simulates a live event without network access. Batch mode reads fixture posts and reveals a few more posts each poll cycle.

Run the first cycle:

APPLE_EVENT_INTERESTS="Siri,Xcode" \
APPLE_EVENT_STORE_PATH=/tmp/apple-event-seen.json \
npm run dev:worker -- --replay --replay-reset --once

Run the next cycle:

APPLE_EVENT_INTERESTS="Siri,Xcode" \
APPLE_EVENT_STORE_PATH=/tmp/apple-event-seen.json \
npm run dev:worker -- --replay --once

Useful replay settings:

  • replay.mode: batch or timed

  • replay.fixturePath: JSON array of LiveblogPost objects

  • replay.batchSize: number of additional fixture posts revealed per poll

  • replay.statePath: JSON file tracking the current replay cycle

  • replay.timeScale: timed replay multiplier, for example 120 means 1 real second equals 2 event minutes

  • --replay-reset: reset the replay cycle counter

Replay still uses the normal seen-post store, so only newly revealed matching posts notify.

Timed replay from captured posts

For a more realistic no-live-event test, capture crawlable WWDC posts locally, then replay their original timestamps against the current clock. Captured publisher content is written under fixtures/local/, which is ignored by git.

npm run capture:replay -- --source all --limit 300 --out fixtures/local/wwdc26-live.replay.json
npm run build

Start the simulation:

APPLE_EVENT_INTERESTS="Siri,Xcode" \
APPLE_EVENT_STORE_PATH=/tmp/apple-event-seen.json \
APPLE_EVENT_REPLAY_STATE_PATH=/tmp/apple-event-replay.json \
APPLE_EVENT_REPLAY_FIXTURE=fixtures/local/wwdc26-live.replay.json \
APPLE_EVENT_REPLAY_TIME_SCALE=120 \
node dist/worker/poller.js --replay --replay-mode timed --replay-reset --once

Run again without --replay-reset to advance the simulated event:

APPLE_EVENT_INTERESTS="Siri,Xcode" \
APPLE_EVENT_STORE_PATH=/tmp/apple-event-seen.json \
APPLE_EVENT_REPLAY_STATE_PATH=/tmp/apple-event-replay.json \
APPLE_EVENT_REPLAY_FIXTURE=fixtures/local/wwdc26-live.replay.json \
APPLE_EVENT_REPLAY_TIME_SCALE=120 \
node dist/worker/poller.js --replay --replay-mode timed --once

Webhook requests are JSON POSTs. If a target has secret, requests include:

x-apple-event-signature: sha256=<hmac>

Environment

  • APPLE_EVENT_INTERESTS: comma-separated interests

  • APPLE_EVENT_POLL_SECONDS: polling interval override

  • APPLE_EVENT_STORE_PATH: JSON dedupe store path

  • APPLE_EVENT_REPLAY_FIXTURE: replay fixture override

  • APPLE_EVENT_REPLAY_MODE: batch or timed

  • APPLE_EVENT_REPLAY_BATCH_SIZE: replay batch size override

  • APPLE_EVENT_REPLAY_STATE_PATH: replay state path override

  • APPLE_EVENT_REPLAY_TIME_SCALE: timed replay speed multiplier

  • APPLE_EVENT_REPLAY_START_AT: fixed simulated start timestamp

  • APPLE_EVENT_REPLAY_ORIGINAL_START_AT: fixed source-event start timestamp

Notes

These sources are public webpages, not stable APIs. Adapters are intentionally small and covered by parser tests so they can be updated quickly when page markup changes.

Related MCP Connectors

  • Dive into the latest and greatest from the tech world with our Hacker News MCP server.

  • Nephia is a brand monitoring service, and this is its remote MCP server. Claude, Cursor, ChatGPT or any MCP client can read the mentions your brand gets on 14 sources: X, Reddit (posts and comments), YouTube, TikTok, Bluesky, Hacker News, Mastodon, Lemmy, GitHub, Product Hunt, Stack Overflow, any RSS feed, Vinted, and AI answers from ChatGPT, Gemini and Perplexity. Every mention arrives already read, with its sentiment and intent, so an agent can answer plain questions: which complaints came in since Friday, what Reddit said about us this week. The source is an argument, not a tool, so one call reads every source you watch. Sign-in is OAuth in the browser: no API key to copy. The consent screen has three permissions: read your mentions and Queries, change what is running (pause, resume, retire), and spend credits (semantic search and AI passes), which arrives unticked. Every tool description states its cost, so a model can budget before it spends. The server is on every plan, Free included, and reading your own mentions through it costs nothing.

  • Hosted MCP server for live public-data APIs and Skills for AI agents.

  • MCP server for AI agents to plan, verify, and deploy Cloudflare-native apps.

Related MCP Servers

  • F
    license
    Not graded
    quality
    D
    maintenance
    An AI-powered news aggregator MCP server that fetches live news from multiple sources, provides AI summaries via Claude, and performs sentiment analysis and trending topic detection.
    -
  • F
    license
    Not graded
    quality
    C
    maintenance
    Local MCP server that ingests and normalizes recent AI, tech, and research news from multiple sources, exposing tools for media agents to query and retrieve article content on demand.
    -
  • A
    license
    Not graded
    quality
    B
    maintenance
    MCP server that aggregates AI news, academic papers from ArXiv, and trending GitHub repositories from multiple sources, enabling users to fetch, search, and filter recent AI content via natural language.
    1
    MIT