watch a changelog (+ webhook delivery)

The watch tool registers change-watch jobs: wigolo snapshots a page, hashes its extracted content, and reports a change the next time the content hash moves — inline, or POSTed to a webhook.

Run it

./watch.sh                                  # watches https://nodejs.org/en/blog
URL=https://your-site/changelog ./watch.sh  # or any changelog you care about

Requires node >= 20 and jq.

What you'll see (real output)

== 1. register a watch on https://nodejs.org/en/blog ==
job id: 74bec5cfdc2cb8bc36a728e46ede89a1b7e7b2a731158771e47674ee6358edfd

== 2. list registered jobs ==
74bec5cfdc2c…  active  every 3600s  https://nodejs.org/en/blog

== 3. on-demand check (first check records the baseline) ==
{
  "url": "https://nodejs.org/en/blog",
  "changed": false,
  "current_hash": "3e600c4fc579"
}

== 4. what a change looks like: the diff engine's report ==
--- old
+++ new
@@ -1,3 +1,4 @@
-## v2.4.0
+## v2.5.0
 - Added container queries guide
+- Added subgrid examples
 - Fixed typo in grid docs

== 5. clean up ==
{"removed_jobs":1}

The first check records a baseline (changed: false + the content hash). Once a check sees a different hash, the report flips to changed: true with previous_hash / current_hash — and for the full picture of what changed, the diff tool renders unified hunks, per-section changes, or a summary (step 4 shows it on two inline changelog versions; wigolo diff <url> compares a page against its cached copy).

Scheduling

A one-shot watch run <id> checks on demand. Scheduled checks fire while a long-lived wigolo is running — wigolo serve or an active MCP session — at each job's --interval. Register jobs once; any long-lived wigolo on the same machine picks them up.

Webhook delivery

Pass --notify at registration and every detected change is POSTed to your endpoint as JSON:

npx wigolo watch add https://nodejs.org/en/blog --interval 3600 \
  --notify https://hooks.example.com/wigolo

{ "job_id": "74bec5cf…", "url": "https://nodejs.org/en/blog",
  "changed": true, "previous_hash": "3e600c…", "current_hash": "9f41aa…",
  "diff_summary": "…" }

A receiver is a few lines of node (run it wherever your automation lives):

// receiver.mjs — node receiver.mjs
import http from 'node:http';
http.createServer((req, res) => {
  let body = '';
  req.on('data', (c) => (body += c));
  req.on('end', () => { console.log('change:', body); res.end('ok'); });
}).listen(8787);

Security note: webhook targets (and watched URLs) must be public http(s) endpoints. Loopback, RFC 1918, link-local, and cloud-metadata addresses are refused by wigolo's SSRF guard — by design, so a watch job can never be used to probe your internal network. Give the receiver a public HTTPS ingress (or whatever tunnel your stack already trusts) and keep redirects in mind: a redirecting webhook target is treated as a delivery failure rather than followed blindly.

Related

• Rich page-version diffs: npx wigolo diff --help
• The same watch tool over REST / MCP for automation platforms: n8n-remote-mcp