Radar (beta)
Watch the conversations shaping your market
Radar is a listening panel inside pendpost. It watches Reddit, Hacker News, Bluesky and Mastodon for people discussing your space, your competitors and the problems you solve, ranks what it finds, and lets you reply behind the same human approval gate as everything else. It is opt-in and off by default: nothing is searched until you enable it per project.
What Radar is for
Five jobs, one panel:
- Show up where your market talks. Join the threads that shape how people pick a tool in your space.
- Build AEO, GEO and SEO relevance. Recurring questions become comparison-page ideas for your site, so AI answers and search results name you.
- Engage communities honestly. Draft a helpful, non-promotional reply and approve it before it posts.
- Watch competitor and comparison talk. See where people weigh you against the alternatives.
- Find real chances to help. Radar ranks signals so you spend time on the ones worth a reply.
How do I set Radar up?
Enable Radar for a project, then give it a first listening query. You have three ways to get there, and you can change any of it later:
- Set it up with your agent. Copy the ready-made prompt into your connected agent. It looks at your product, proposes a query, and saves it for you over MCP with
config_set. - Add one by hand. Fill in a short form: a label, the keywords your buyers and community use, competitors, and the best subreddits or hashtags.
- Let your agent research your space. Paste the research prompt and your agent studies your market first, then proposes queries from what it finds.
Every Radar action a person can take in the panel can also be driven headlessly over MCP, so an agent can enable Radar, manage queries, scan, triage a signal, or queue a reply.
Manual or daily?
Scan a query by hand whenever you want, or mark it Daily. A daily query is scanned once every 24 hours and folded into your morning digest, so fresh signals reach you without you opening the panel. There is no cadence or rate-limit engine behind it, just a once-a-day sweep.
Replies stay behind the approval gate
A Radar reply never posts on its own. You (or your agent) draft it, it runs through the humanizer so it reads human, and it enters the normal queue as a draft. A distinct person approves it before it publishes, and it can never auto-approve. That is the same fence pendpost applies to every post: replying into someone else's thread is exactly the action that should always have a human on the trigger.
Reddit: you approve, then it posts
Reddit posts publish through pendpost's account-warmth screening. A distinct person still approves every Reddit post; nothing auto-approves. After that approval, the post publishes on its own, whether it is organic or promotional. If your account is new or low on karma, or a post looks promotional, Radar shows a "Heads-up" advisory on the approval card so you know before you approve, but the post still publishes once you do. Warmth screening is a warning you can see, not a hidden gate.
Get found in AI answers
When people ask AI tools about your space, the same questions come up again and again. Radar collects those into comparison-page ideas you can turn into pages on your site, and it tracks whether the models actually name you. Over time you write the pages that answer the real questions, and you can watch your mention rate move.
Full setup, including platform credentials, is in the documentation.