Heardley, a Reddit lead generation and monitoring tool.
500
Reddit
alerts,
down
to
the
5 that can buy.
Be heard. Be cited.
Heardley is a Reddit lead generation and monitoring tool. It scores every thread for buyer intent, drafts the reply in YOUR voice, and scrubs the AI tells before you post. For founders and marketers who grow on Reddit but don't have three hours a day to doomscroll for the one thread that matters.
500 Reddit threads in. 5 worth a reply out.
That's the whole product. Intent-scored, AI-tell-linted, delivered as a digest at 9am.
Pay once, paste a free key, start finding leads.
Cheaper than a single month of GummySearch's old Pro plan. You're paying for the software, not the AI calls. Hosted, monitored, and updated by us so you never touch a server.
See what's included →Sign in with Google or GitHub, hit Create Key, copy the value. The default model is free for personal use. Heavy users spend $2-$5/mo in tokens, billed by OpenRouter, not us.
How do I get one? →Paste your key into the yellow banner at the top of the app. Click Scrape now. The triaged digest lands in your inbox at 9am the next day, ranked by buying intent.
Start free →The tools you used last year aren't coming back.
GummySearch and Howitzer both shut down in late 2025. The replacements are scrambling. Their old customers are stuck searching for something that works the same way at a price that isn't insulting.
You don't need more alerts. You need fewer, better ones.
Every Reddit tool floods your inbox. The good threads are buried in 95% noise, and the cost of reading through them is two hours a day you'll never get back.
Keyword alerts don't equal buyers
F5Bot sends you everything that mentions your category. Most of it is people venting. Some of it is bots. Almost none of it is someone holding a credit card.
Reply too generic and you get dunked on
Every "Great question! In today's competitive landscape..." reply ends up at the bottom of the thread, downvoted into invisibility. Worst case, the account gets shadowbanned.
Sub-$99 tools are blunt, $99+ tools are bloated
Either you get a noisy keyword feed for $19 or a multi-platform suite that buries Reddit under five other channels. Neither one is actually built for finding the next ten customers.
Every thread gets scored before it touches your inbox.
Our scorer reads the title, body, and the surrounding context, then classifies the thread by intent (3 = looking for a tool, 2 = switching from a competitor, 1 = casual problem talk, 0 = noise). Only 2s and 3s make the digest.
The dashboard that filters Reddit's firehose into 5 threads a day.
Hook up your OpenRouter key (free tier works) and we'll score every scraped thread on a 0-3 intent scale, color-coded by category. You see only the threads where someone is actively asking for what you sell. There's a one-line angle next to each thread suggesting how to reply.
- Intent classification on every thread (3 = high, 0 = noise)
- Category buckets: solution request, switching, pain point, comparison
- One-line "angle" suggesting how to reply (not the reply itself)
- Daily email digest at the time you set, ranked by score
Replies that sound like YOU, not a chatbot.
Every other Reddit tool generates the same reply: em-dashes, "leverage," "Hope this helps!" Reddit users spot it in half a second and your account gets shadowbanned. The fix isn't a better prompt. It's showing the model who is supposed to be writing.
Build a persona (22yo Linux gamedev who drops articles and uses "fr"; or 35yo SaaS founder who writes in clean paragraphs and references their cofounder). Paste 2 to 5 of your favorite Reddit comments as style examples. Then on every thread, click Draft reply and watch a draft come back in that voice. Run it through the linter for the final pass and copy.
- 200+ persona dimensions: age, gender, region, voice, formatting, quirks
- Few-shot style examples (paste a Reddit URL, we fetch it)
- Multiple personas per account so each sub sees a different "you"
- Auto-scrubbed for em-dashes, AI openers, and the rest before you see it
And if you wrote the reply yourself? Paste it, we still lint.
Draft Reply produces the polished version. The linter is the standalone tool you reach for when you typed something yourself, or when you're auditing a reply your VA wrote, or when ChatGPT gave you a draft and you want to know how shadowban-shaped it looks.
It catches 50+ AI-tell patterns deterministically (em-dashes, banned-word vocabulary, parallel bullets, "Hope this helps!", pitching too early) and gives you a one-sentence biggest-fix. Verdict: post it, tweak it, rewrite it.
- Catches 50+ AI-tell patterns deterministically (no LLM needed)
- Optional second-pass AI voice check (specificity, vendor energy)
- "Biggest fix" sentence so you know exactly what to change
- Verdict: post it, tweak it, or rewrite it
Paste a competitor URL. Get back what your sales team should be saying.
The Intelligence module fetches the competitor's landing page, extracts their positioning, then searches Reddit for what users actually say about them. You get a counter-positioning brief plus the keywords and subreddits to start tracking.
We pointed it at Mailchimp. The tool came back with the top Reddit complaints (forced tier jumps for one feature, contact-count billing surprises), suggested keywords like "affordable Mailchimp alternative", and a counter-positioning angle a marketing consultant would charge $500 for.
- Positioning extraction (value prop, ICP, pricing tier, tone)
- Reddit sentiment with real quotes from users
- Counter-positioning advice you can use in your replies
- One-click "apply suggestions" to add keywords/subs to your project
Mailchimp analyzed
All-in-one email + marketing automation for SMBs. Positions on ease-of-use and the "free up to 500 contacts" tier as the on-ramp.
Reddit sentiment · 184 mentionsPosition on transparent flat pricing: one tier, every feature, contacts scale linearly. Speak to founders who got burned by a 3x bill jump after one bad month.
15 minutes from "what's a YAML file" to "I found my first buyer."
Paste your URL or pitch
The onboarding wizard reads your site (or your one-paragraph description) and writes your project for you: keywords, subs, competitors, negatives. Edit and save.
Add your OpenRouter key
One field in Settings. Free tier works. Every thread you score costs you about $0.02 in tokens, billed by OpenRouter, not us. We never see the spend.
Hit "Scrape now" and read your inbox
The scraper pulls Reddit, the scorer triages, the digest lands at 9 a.m. tomorrow. You write 1-3 replies in 15 minutes, lint each one, and ship.
Solo for founders. Team for shared work.
Keep personal projects private, or share exactly one Heardley project with coworkers and clients. Team plans unlock project sharing inside Heardley.
- Unlimited solo projects + scoring
- Reply linter, intent buckets, switch radar
- Competitor Intelligence module
- Daily email digest
- Hosted dashboard at heardley.com, signed in from any device
- Everything in Solo Monthly
- Every future update at no extra cost
- Priority support straight from the dev
- Founding-customer badge in-app
- Hosted SaaS, signed in from any device, zero setup
- Up to 5 seats
- 5 shared Heardley projects
- Shared leads, personas, intel, keywords, and knowledge bank
- Viewer, member, and admin roles
- Up to 15 seats
- 25 shared Heardley projects
- Keep every client project isolated
- Enterprise available for custom limits
Questions you'd ask before paying.
Why is this so cheap?
Because you bring your own OpenRouter key. We're charging for the software, not the AI tokens. Our hosting bill per user is under a dollar a month. $9 covers it, $54 covers it for years.
Do I have to pay for AI tokens on top of the $9?
No. OpenRouter offers a free model that powers every Heardley AI feature at $0. You sign up for OpenRouter (free, no credit card, 2 minutes), paste the key into Settings, done. If you ever want a smarter paid model the typical heavy-user cost is $2-$5/mo billed by OpenRouter directly. Other tools bake $20-$60 of LLM markup into their subscription; we don't.
Will it get my Reddit account banned?
We never post anything. We only read public threads, same as someone scrolling Reddit. The reply linter exists specifically to keep your account safe by catching the AI-tell patterns that get flagged.
What if you shut down like GummySearch did?
Everything you put into Heardley is exportable as CSV/JSON from Settings. If we ever wind down, we'll give 90 days' notice and a one-click full data export. Your data isn't locked behind us.
Does it really replace GummySearch?
For audience research, subreddit discovery, and intent tracking: yes. We don't have the historical archive GummySearch built over 7 years. For live monitoring and lead discovery, the same workflow is here, often with better filtering.
Can I run more than one company?
Yes. Add as many projects as you want, different products, different brands, agency clients. There's no per-project fee.
Does it work for non-English subreddits?
Yes, as long as your OpenRouter model handles the language. We've tested Spanish and German and they score about as well as English.
Stop reading 500 Reddit posts. Start replying to 5.
Lifetime is $54. That's lower than a single month of every dead competitor. Worth the gamble.
Get your OpenRouter key in 4 steps
Sign up at OpenRouter
Go to openrouter.ai and sign in with Google or GitHub. No credit card required.
Create a key
Open the Keys page, click Create Key, name it "Heardley", and copy the value (starts with sk-or-v1-...).
Paste it into Heardley
Open the app, hit the yellow banner at the top, paste your key, click Save. The app validates the key against OpenRouter before saving.
Scrape and score
The default model is openai/gpt-oss-120b:free, it's free for personal use. Heavy users typically spend $2-$5/month. Switch models any time in Settings.