How a page change becomes an alert you can act on
No black box. Here is the whole pipeline, from the portal we fetch to the email in your inbox, including what we deliberately do not do.
Four stages, running around the clock
Fetch
Scheduled fetchers pull pages, PDFs, feeds, and portal results. Raw snapshots are stored, so we can always answer "what did it say before".
Diff
Hashes, then semantic diffs for tables and documents. We see exactly what changed, not just that a page changed.
Triage
The change is scored against your filters and a written relevance rubric. Low-confidence items go to a human review queue, not to your inbox.
Deliver
You get the change, the diff, the source link, and context. Digest by default, instant on Growth and Pro, Slack and webhooks on Pro.
Your first week
Setup is a form, not a project. Most customers are receiving alerts within 48 hours.
You tell us what matters
Pick a monitor. Give us your states, keywords, brands, domains, or case types. Five minutes, no calls required.
Sources get wired
We add the public sources that match your scope. If something you asked for is not feasible, you hear it from us immediately, with alternatives.
First digest
You receive a baseline digest showing what exists right now, so future alerts read as changes instead of a wall of links.
Tuning
Too noisy? Too quiet? Both are fixed with a filter change, not a support ticket that disappears into a queue.
It just runs
Source health is monitored with canary checks, so a parser silently breaking gets caught by us, not by you noticing an empty week.
Silent failure is the only enemy
Canary checks
Every source is probed on a schedule. If a fetch pattern breaks, we know before the silence reaches your inbox, and we fix it.
Human review queue
Ambiguous changes never reach customers unreviewed. A person looks at anything below the confidence threshold before it ships.
Adjustable precision
You set the tradeoff: tell me everything, or only high-confidence changes. One setting, applied to every source in your scope.
What we do not do
The limits are as important as the features. If your use case needs something on this list, we are the wrong vendor, and we will tell you so.
- We do not log into systems that are not yours. Public sources only, unless you explicitly authorize us in writing to monitor your own accounts.
- We do not give legal, financial, or compliance advice. We surface changes and link to the source. Decisions are yours, or your attorney's.
- We do not file documents for you. Where a platform process applies, you get the evidence pack and the filing path, and you decide.
- We do not sell or resell your data. Watchlists, keywords, and alert history are yours. We sell monitoring, nothing else.
Where alerts land
- Email digest (every plan)
- Instant email alerts (Growth and Pro)
- Slack, Telegram, webhooks (Pro)
- API and MCP for your own agents (Pro)
Security and privacy
- Encryption in transit and at rest
- Least-privilege access, audit logs
- Written authorization required for security monitoring targets
- Delete your data on request, no dark patterns
See it on your own sources
The fastest way to judge this is your own watchlist. Start the trial, point it at the pages you check by hand, and compare notes after a week.