I built Guru Tracker to expose fake financial gurus — $4.2M losses prevented
Hey Indie Hackers 👋
I'm launching Guru Tracker today, and I wanted to share the story
behind why I built it.
THE ORIGIN STORY:
I lost $3K following a "top crypto trader" on Twitter. When his
predictions failed, he simply deleted the tweets. His track record
disappeared. My money didn't.
That's when I realized: there's zero accountability in the financial
influencer space.
THE PROBLEM:
- 67% of "experts" delete failed predictions within 24 hours
- Average prediction accuracy: 12.4% (worse than random)
- Retail investors lose billions annually
- No way to verify a guru's real track record
THE SOLUTION:
We built Guru Tracker to permanently record and verify financial
predictions using:
✅ Forensic OCR (recover deleted tweets)
✅ AI-powered analysis
✅ Real-time market verification
✅ Immutable leaderboards
CURRENT METRICS:
- 4,203 predictions tracked
- 892 experts exposed
- Average accuracy: 12.4%
- $4.2M losses prevented from our community
BUSINESS MODEL:
Free tier + Premium ($16/mo) + Enterprise for trading communities
THE ASK:
We're indie hackers too — no VC funding, bootstrapped, mission-driven.
Before you follow any financial influencer, check their real record.
What do you think? Would this be valuable in your investing journey?
the delete-and-disappear move being the core wedge is a sharp catch, that's exactly what makes these people slippery. $4.2M is a strong number if it holds. the thing i'd think hard about is the legal surface, calling specific people fake publicly invites pushback fast, a track-record-receipts framing that lets people judge for themselves feels safer than 'exposing' them tbh
his is good advice and it's actually how we built it. we don't call anyone a fraud we just publish their record. date, claim, outcome. win or loss. the user decides. "guru X said BTC hits 10k by March, here's what happened" is a fact, not an accusation. the legal surface shrinks a lot when you're just a scoreboard. the framing you're describing is exactly where we landed.
smart, the receipts-not-accusations framing is what keeps you out of court and honestly what makes it land harder too, a clean dated record reads worse than any "fraud" label ever could. i'd put that claim log front and center, it's the thing people screenshot. rooting for this one.
Exactly this. The claim log is already the centerpiece — every prediction gets a permanent entry: date, exact claim, outcome, whether it was deleted or not. The deletion flag is actually what people screenshot the most. There's something viscerally satisfying about seeing "deleted 3 days after the deadline passed." Thanks for rooting for this — if you want to see it in action, guru-tracker.com has the live leaderboard.
product makes instant sense which is rare. my one worry is the data, these gurus are spread across tiktok/youtube/telegram/discord and the receipts get scrubbed fast. the moat is archiving their claims in real time before they delete. the tracking ui is the easy bit. if you're pulling public stuff after the fact you're always a step behind the scrub ngl. how are you capturing it, manual or automated?
real-time capture, not retroactive pulls. automated across platforms, timestamped at publish. gurus can delete we don't care. happy to go deeper on the architecture if you're curious.
ok that's exactly the answer i was hoping for, real-time timestamped capture is the whole moat and everything else is replaceable. one gap i'd keep an eye on is the stuff that never hits public feeds, the paid telegram/discord pump rooms where the worst calls actually live. if you can get receipts out of even a couple of those it gets a lot scarier for them. architecture sounds right though.
This hits a very real problem — especially the “delete and disappear” behavior.
That alone makes the product instantly understandable.
A couple thoughts to make this even stronger:
→ Trust layer will make or break this
People will ask:
how accurate is the OCR recovery?
can gurus dispute entries?
how do you prevent false positives?
If you can show proof here (examples, audit logs), it becomes much more credible.
→ Positioning opportunity
Right now it feels like “exposing bad actors”
But it could also be:
→ “find who’s actually worth following”
That’s a more positive + broader use case.
→ Distribution angle
This feels very shareable if you lean into:
“Top 10 worst traders this week”
“Most deleted predictions”
“Real accuracy leaderboard”
People will naturally circulate that.
→ Big unlock
If you can show:
→ “following X vs random vs market” performance
that becomes insanely compelling.
Curious — have any influencers pushed back yet or tried to dispute their scores?
Also, I’m running a small project (Tokyo Lore) where we test tools like this with a focused group of builders and see how they actually react (especially trust-heavy products like yours).
Since you’re tackling accountability + real-world impact, this could be a strong fit to validate how people respond beyond initial curiosity.
Happy to share more if you’re interested 👍
this is genuinely useful, thank you. the trust layer point is the one we think about most OCR accuracy, dispute process, audit trails. we're building toward public methodology docs so anyone can challenge an entry. false positives are the thing that kills credibility fast, we know that.
the "find who's worth following" framing is something we've been debating internally it's a better front door for people who aren't already burned. leaderboard-style distribution is already on the roadmap, "most deleted predictions" is exactly the kind of thing that spreads without us pushing it.
on disputes not yet at scale, but we expect it. the answer is always the same: here's the original timestamped claim, here's the outcome. hard to argue with a screenshot and a date.
on Tokyo Lore appreciate the offer but we're heads down on the core build right now. happy to revisit when we're further along.