Urgency on every mention
Claude judges how soon to respond. High means an outage many people hit, a false claim spreading, or an angry customer with a large audience. Filter to Urgent only.
Reputation monitoring
Catch the complaint before it becomes the thread.
A bug many people hit, an angry customer with a large audience, a post with the wrong price: Overhear finds them every hour and flags the ones that need you now.
You hear last
Support tickets arrive after the public post, if at all. By then the thread has its own version of events.
False claims stick
Wrong pricing or a missing feature, repeated in a popular thread, becomes what people believe about you.
Not everything is urgent
A feed that treats every mention the same buries the one post that needed an answer this morning.
How Overhear does it
Misinformation
Posts that state something false or misleading about you, like wrong pricing or features, get their own type and always trigger an alert. Claude judges claims against the facts you enter on Settings: pricing, plans, platforms, what you do not do.
Fact check
Every week Overhear reads the pages that describe your product ("vs" pages, alternatives lists, pricing pages, app listings) and checks them against your confirmed facts. A page with your old price or a feature you never had becomes a misinformation mention, with the quote and the fact it gets wrong.
Complaints and bug reports
Frustration and concrete defect reports are typed separately, with the sentiment toward your product recorded on each mention.
One email per scan
Alerts go to every member of your organization after a scan, only with the mentions that need action. No news, no email.
Webhook
Send every new mention, with its type and urgency, to your own endpoint as a signed POST.
What it costs
One brand keyword on all six sources at the default cadence: about 2640 credits a month, $4.40 at the $10 pack rate.
Prepaid credits, no subscription, no keyword cap. Credits never expire and a failed search costs nothing.
Questions
Does Overhear reply for me?
No. It drafts a reply you can edit and copy; you post it yourself, so nothing goes out under your name without you.
Who receives the alerts?
Every member of the project's organization. Alerts are off by default and turned on per project under Settings.
Which mentions trigger an alert?
High urgency, misinformation, buying intent, and leads with active intent and high fit. Mentions judged unrelated never do.
More use cases
Brand monitoring
Track brand mentions on X, Reddit, LinkedIn, YouTube, Hacker News and the web. Claude sorts each one into praise, complaints, questions and buying intent.
Competitor monitoring
Track competitor mentions on X, Reddit, LinkedIn, YouTube, Hacker News and the web. Find complaints about rivals and people asking for an alternative.
Lead finding
Find people on Reddit, X, LinkedIn, YouTube, Hacker News and the web who ask for a tool like yours, scored for fit and buying intent, with drafted replies.
Wall of love
Build a wall of love from posts people already wrote. Overhear finds praise on X, Reddit and more, and your site reads what you save from a JSON feed.
Reddit monitoring
Monitor Reddit for your brand, competitors and buying-intent phrases. Hourly keyword scans, AI-sorted mentions and email alerts, 1 credit per scan.
X (Twitter) monitoring
Monitor X (Twitter) for posts about your product or competitors, even when nobody tags you. Hourly scans, sorted by type and urgency, 2 credits a scan.