Scrape → Analyze → [Decide]

Stop pulling data.
Start pulling verdicts.

Type the decision you're trying to make. The agent picks sources, scrapes them, runs the analysis, and gives you a one-line answer with the receipts.

heeo · session 0xA9F3 · live
should I long $SOL into the Jan unlock?
scanning twitter… 412 accounts
pulling on-chain unlock schedule…
checking funding rates…
cross-referencing sentiment vs price action…
Wait. Sentiment is leading price by 6 days — short window opens Jan 14.
// try it

Type a decision you're avoiding.

Pick a starter or write your own. We'll show you what the agent would scrape, what it would analyze, and what it would tell you to do. No signup until you want the full report.

HEEO_AGENT · v0.4.2 ● ready
which 3 competitors are quietly winning my niche? is this token's twitter sentiment fake? should I sponsor this newsletter? is this YC company hiring fast enough to be real?
> type the decision you need to make RUN →
// the gap

Apify gives you a CSV. We give you the call.

Same scrape job. Same 412 twitter accounts. The difference is what you walk away with.

WHAT YOU GET FROM A SCRAPER
handlefollowerstweettimestampengagement
@solwhale_14,221SOL is going to $400 EOY trust me2024-12-18 03:141,204
@degen_alpha9892massive unlock incoming, accumulate2024-12-18 03:2187
@cryptoking423,109nobody is bullish enough on $SOL2024-12-18 04:02412
@onchain_pepe619jan unlock is bullish actually2024-12-18 04:1852
@yieldfarmer__2,440SOL setup looking clean here2024-12-18 04:33198
@trader_anon11,802longing into the unlock, easy money2024-12-18 05:01933
@flipperking402load up on $SOL before jan 142024-12-18 05:1423
@degentrader001,221SOL bull case is undeniable2024-12-18 05:39114
@web3frog5,028this dip is a gift, longing2024-12-18 06:02302
@chartwizard_8,191SOL/USD breakout imminent2024-12-18 06:18551
Raw rows
412
Time to insight
~4 hours
WHAT YOU GET FROM HEEO
64% of the bullish tweets came from accounts created in the last 90 days. Sentiment is paid.
  • 264 of 412 bullish accounts < 90 days old → twitter.com/i/...
  • Wallet cluster overlap: 4 funders → dune.com/heeo/cluster
  • Engagement-to-follower ratio 8.2× organic baseline → internal
Skip the long. Re-evaluate after Jan 14 unlock.
Verdict
1 sentence
Time to insight
~90 seconds
// real runs

Three real runs from last week.

Anonymized but actual. Each is a question someone asked the agent, the sources it pulled, and the verdict it returned.

TRADER · 11:42PM TUE
is the bullish sentiment on $WIF organic or coordinated?
twitter dune discord
Coordinated. 71% of top-engagement replies trace to 4 wallet clusters.
  • Cluster A funded 38 accounts via same exchange withdrawal pattern
  • Top 200 replies share 92% lexical overlap (n-gram analysis)
  • Discord moderators cross-posted same talking points within 4-min window
INDIE FOUNDER · 9:15AM MON
which 3 competitors in 'AI sales tools' are growing without raising?
linkedin producthunt wellfound github
Two are quietly profitable. One is faking traction with bot reviews.
  • Company A: 14→47 employees in 11 mo, no funding round filed → linkedin
  • Company B: 3 founders, $0 raised, 2.4k MAU on PH listing → producthunt
  • Company C: 89% of reviews from accounts <30 days old → flagged
ANALYST · 4:08PM FRI
is this Series A startup's hiring pace consistent with their claimed ARR?
linkedin wellfound crunchbase
No. Hiring froze 6 weeks ago. ARR claim is likely 90 days stale.
  • 0 new hires posted since Nov 2 (vs 4-5/mo prior trailing 6mo)
  • 2 senior eng departures in last 30 days, not backfilled → linkedin
  • Wellfound listings reduced from 12 → 3 active roles
// spec sheet

What's under the hood. No hype.

We don't do magic. We do clean scrapers, source-aware models, and a refusal layer that flags low-confidence answers instead of bluffing them.

SOURCES SUPPORTED
Twitter, LinkedIn, Reddit, Discord, on-chain, GitHub, ProductHunt, Wellfound, Crunchbase, custom
SCRAPE THROUGHPUT
~2,400 rows/min per job
ANALYSIS MODELS
Claude 3.5 Sonnet + a fine-tuned classifier for source-trust scoring
CONFIDENCE FLOOR
0.72 — below this, the agent refuses to give a verdict
RECEIPT REQUIREMENT
Every claim must cite ≥2 source rows
REFUSES WHEN
Source pool is too thin, sentiment is paid, or the question is unanswerable from public data
"If we can't source it, we won't say it."
// pricing

Pay for verdicts, not seats.

Two ways to use it. Pick the one that matches how you actually work.

PER DECISION
$0.20 / report
You ask a question every now and then
  • Up to 5,000 rows scraped per run
  • All sources
  • Receipts + source links
  • No subscription
Run one →
OPERATOR
$29 / mo
You run decisions weekly and want them saved
  • 100 reports/mo included, then $0.04/row
  • Saved scrapers + scheduled runs
  • Slack + email delivery
  • API access
Get Operator →
No free trial. First decision is $0.20 — run it before you commit.
// a note

A note on why this exists.

I built scrapers for traders for two years. Every single one of them asked the same follow-up: "okay but what should I do with this?" Apify gives you the rows. ChatGPT will hallucinate an answer. Neither will tell you when the data isn't enough to decide. That gap is what this is.

If you're already comfortable wiring up scrapers and prompts yourself, you don't need me. If you'd rather type the question and get the call, that's what you're paying for.

— founder, building heeo