The conversations that matter most are happening without you.
Right now someone is asking for recommendations in your category, someone is venting about your competitor, and a customer is airing a problem you could fix today.
ThreadSonar is social listening with judgment: it watches X, Reddit, LinkedIn, Threads, Bluesky, and Hacker News around the clock, works out which conversations matter and who is talking, and maps them into one inbox, your Slack, or your AI agent over MCP.
leads marked not-relevant are credited back. you pay for qualified, never for junk
sonar contacts · 🦞 = a conversation that matters
sweeping…
Every thread does two jobs.
Today, that “what should I use for X” thread is a live buyer you could win. For the next three years, it's a source ChatGPT and Perplexity will cite when every other buyer asks the same question.
Join the conversation, win the person in it
Reply while the thread is live: win the buyer asking, save the customer venting, correct the record on your product. Real outcomes this week, from conversations already happening.
Be present in the threads AI trusts
AI engines learn who to recommend from these exact threads. Show up in them and you get recommended to buyers you'll never talk to. That's your share of model.
the gap
Almost nobody does either. Not because they can't write a reply, but because they can't find the threads: millions of conversations, and the relevant ones die within hours.
ThreadSonar finds them. It doesn't post, doesn't send, doesn't automate outreach. It answers one question extremely well:
Where should you be talking right now?
Buyers now ask AI first. Are you in the answer?
That headline up top means these conversations too. When someone asks ChatGPT or Perplexity “what should I use for X,” a shortlist comes back, a vendor gets chosen, and no dashboard anywhere records that it happened. It is the fastest-growing part of your funnel and it is completely invisible.
ThreadSonar makes it a number. We ask your category's buying questions to every major engine on a schedule, track the share of answers that include you, and show which competitors own the rest.
Then it makes the number movable. Engines learn who to recommend from public threads, and they cite their sources. We surface the exact threads behind the answers that skipped you, and those threads are listenable: one click turns them into a signal, and showing up in them is how the number changes.
illustrative data · the real board lives at /geo in your workspace
Triage in one tap. It learns from every tap.
This is the real inbox, so click around. 👍 tunes your signal sharper, 👎 credits the lead back and asks why, and new leads drop in the moment they qualify.
- 95HOTSarah ChenFounder @ PropFlow · CRM for property managers
Since Intently shut down we've been scrambling for a replacement. Need a social listening tool that watches Reddit + X for people actually asking for products like ours. Happy to pay $200-300/mo if the leads are qualified. What is everyone using?
Founder explicitly seeks a replacement after losing access, budget and criteria stated.· Intently refugees
Open post - 91HOTPriya NatarajanVP Marketing @ Craftline · B2B manufacturing SaaS
We're consolidating our GTM stack next quarter and social listening is the biggest gap. If your team found something that surfaces actual buying conversations, not vanity mentions, I'd genuinely love an intro. Budget is approved.
Approved budget + active vendor search posted to her network.· Solution seekers
Open post - 88HOTgtm_danRuns listening for 6 B2B clientsr/marketingagency
Our agency runs listening for 6 B2B clients and Brand24 is getting way too expensive for what it delivers. It keeps flagging junk and missing the actual buying conversations. Is there anything that scores posts for intent?
Agency operator complaining about competitor precision and price. Ideal ICP.· Competitor complaints
Open post - 72WARMmkozlowskiBuilding devtools. Prev: infra @ StripeAsk HN
Ask HN: how do you find early users for a niche B2B product? I've been manually searching Reddit for people complaining about the problem we solve. It works but takes ~5 hrs/week and I definitely miss threads.
Describes the exact manual workflow the product automates, with quantified pain.· Solution seekers
Open post
Marketing & GEO
Brand and competitor mentions, sentiment, share of voice, and the dark-social threads where your category gets discussed. The same threads are what AI engines cite, so showing up in them grows your share of model too.
Product & Customer Success
These threads are a gold mine of issues: what people hate about your competitors, what frustrates them about you, and your own customers venting before they churn. Straight into the roadmap, or a save.
Sales & GTM
Founders, SDRs, and GTM engineers: catch buying signals in the wild. Pain, asks for recs, competitor complaints. Reply while the intent is hot, straight into pipeline.
Say it in plain English
“Agencies complaining their social listening tool is too expensive.” That's the whole setup. We compile it into platform-native queries, a scoring rubric, and semantic anchors.
Every match is judged
A three-tier funnel drops the junk before it costs anything: hygiene checks, a semantic gate, then an LLM judge that scores 0–100 for relevance and urgency, tags who's talking (a buyer, your customer, a feature request, a competitor's user), and attaches the reason.
It gets sharper weekly
One-tap 👍/👎 on each lead tunes your signal's precision automatically, and anything you mark not-relevant is credited back.
One sentence away from any audience.
Signals are plain English. If you can describe the conversation, we can put it on your radar. Each of these is one click from being yours.
“Alert me the moment someone asks how to cancel our competitor”
Salescreate ↗“Catch our own customers venting publicly before they churn”
Customer successcreate ↗“Find every 'what do you all use for…' thread in our niche”
Demand gencreate ↗“Find users rage-posting about a competitor's price hike”
Competitivecreate ↗“Show me strangers ranting about the problem we solve”
Founderscreate ↗“Catch job posts that reveal a company will need us next quarter”
RevOpscreate ↗
Signal data your team, or your agents, can act on.
Live signals are only worth what you do with them. Route them to a Slack channel for humans, or plug any agent straight into the data over MCP.
Slack, for your team
Every qualified lead drops into the channel you choose, with the post, the score, the why, and a reply draft. Your team sees it while the thread is still warm and answers in a tap.
Each workspace connects to its own Slack, even a different Slack account per product.
An MCP server, for your agents
ThreadSonar is a live conversation-signal data layer any agent can tap: point Claude, Cursor, or your own agent at our MCP server with a workspace API key and it can pull fresh hot leads, use the reply drafts, and mark each one contacted.
Listening to acting, no human in the loop required.
fluent in the agent ecosystem openclaw kicked off 🦞
> get_leads { status: "new", band: "hot" }
{ "count": 3, "leads": [ { "intent_score": 88,
"rationale": "asking for alternatives to…",
"reply_draft": "Sounds rough. We…" } … ] }
> mark_lead { status: "contacted" }connect an agent →Share of voice in threads
Every listened post tagged by who got named: you or a competitor. The thread-level twin of the share-of-model board.
Engager mining
The people liking and replying to competitor posts become a warm audience list you can export.
Account watch
Follow a competitor account and its execs. Every post, every engager, every slip.
Verified contacts
Qualified leads arrive with verified work email and mobile, resolved through a multi-vendor waterfall.
All six live today. Every post scored by the same funnel, delivered to the same queue.
Someone just posted about the problem you solve. Someone else is venting about your competitor.
People call this category social listening. That undersells it: listening is passive consumption. This is knowing which conversations to be in, before they're over.