Anthropic Blog Tracker

What does Anthropic Blog Tracker do?

Anthropic Blog Tracker turns the official Anthropic news page into a clean, queryable feed. Every post is fetched, classified by topic (model release / safety / research / business / customer story / policy / announcement), enriched with authors, tags, an excerpt, and emitted as a structured JSON row. Use it once or schedule it on the Apify platform to get a daily diff of what Anthropic just announced — with API access, monitoring, retries, and proxy rotation included.

Built by Ian Mu as part of a 100-Apify-Actor portfolio. Companion repo: github.com/ianymu/claude-verify-before-stop — a Claude Code hook that enforces "verify before stop" on agent runs.

Why use Anthropic Blog Tracker?

• AI watchers / journalists: get a structured feed of Anthropic's announcements instead of refreshing the news page.
• Competitive intel teams: filter to model-release or business and pipe into a Slack channel or BI tool.
• Founders / PMs: track when new Claude models or capabilities ship — Opus, Sonnet, Haiku, tool use, context window changes — and react fast.
• Researchers: filter to safety and research for a clean reading list.
• Aggregators / newsletter writers: drop the JSON into your own template, no manual copy-paste.

Compared to RSS, this Actor adds topic classification, author/tag enrichment, and filterable date ranges, all without you owning a server.

How to use Anthropic Blog Tracker

1. Click Try for free to open the Actor.
2. Optionally narrow the run with Since date (default: last 90 days) and Filter by categories.
3. Set Max results if you only want the top N.
4. Click Start. The run finishes in well under a minute for default settings.
5. Open the Dataset tab and export as JSON, CSV, Excel, or HTML — or fetch via the Apify Dataset API.
6. To run on a schedule, click Actions → Schedule and pick a cadence.

Input

Field	Type	Default	Description
sinceDate	ISO date string	90 days ago	Lower bound on publishedAt.
categoryFilter	array of strings	[] (all)	One or more of model-release, safety, research, business, customer-story, policy, announcement, other.
maxResults	integer	50	Hard cap on emitted rows.
fetchBody	boolean	true	If false, skip the detail-page fetch (faster, fewer fields).

Example input:

{
  "sinceDate": "2026-01-01",
  "categoryFilter": ["model-release", "safety"],
  "maxResults": 20,
  "fetchBody": true
}

Output

You can download the dataset in various formats such as JSON, HTML, CSV, or Excel.

{
  "title": "Claude 4 Opus 1M Context",
  "url": "https://www.anthropic.com/news/claude-4-opus-1m-context",
  "publishedAt": "2026-05-15",
  "category": "model-release",
  "authors": ["Anthropic"],
  "tags": ["claude-4", "opus", "context-window"],
  "excerpt": "Today we're introducing Claude 4 Opus with a 1M-token context window...",
  "bodySize": 4231
}

Data table

Field	Type	Notes
title	string	Post title (og:title preferred).
url	string	Canonical anthropic.com URL.
publishedAt	string (YYYY-MM-DD)	From article:published_time or <time> element.
category	string	One of 8 categories (see input).
authors	array of strings	From meta[name=author] / article:author.
tags	array of strings	Meta keywords + topic links + auto-tags.
excerpt	string	First ~300 chars of body.
bodySize	integer	Total length of extracted body text.

Pricing / Cost estimation

This Actor runs on plain HTTP + Cheerio — no headless browser, no proxy needed. A typical run touches one index page + up to maxResults detail pages at 1 req/sec (respectful pacing). Most runs finish in 0.01–0.03 compute units, well within the free Apify tier. Scheduling daily for a year typically stays under $1 of platform cost.

Tips or Advanced options

• Set fetchBody: false if you only need title + URL + date + category — the actor skips the detail fetch and runs in seconds.
• Combine categoryFilter with sinceDate to feed exactly the right rows into a downstream Zapier / Make / n8n workflow.
• Run hourly with sinceDate set to 1 day ago for a low-latency announcement watcher.
• The classifier is heuristic — if it mislabels a post, open an issue and we'll tune the rules.

FAQ, disclaimers, and support

• Is this affiliated with Anthropic? No. This Actor scrapes the public news page and is provided as-is.
• Is this legal? Scraping public, non-personal blog data is generally legal; respect Anthropic's terms and Robots.txt. The actor uses a clearly-identifying User-Agent and 1 req/sec pacing.
• Why is a post missing? Anthropic occasionally moves or unlists posts; the index page is the source of truth, so anything not linked there won't be picked up.
• Bug or feature request? Open an issue on the GitHub repo, or contact Ian Mu via github.com/ianymu. Custom variations (e.g. push directly to Postgres, classify with an LLM, alert on keyword) are available.
• Related: see claude-code-changelog-watcher — same idea but for the anthropics/claude-code repo.