Ocellus @ocellusapp

Exits now come to your Telegram. I’ve added the Exits feed to @OcellusBot. Add it to your chat and type /exits. You’ll get alerts when a creator sells or liquidity leaves the launchpad’s locker on a scanned Robinhood Chain coin. Each alert shows how much moved, the coin’s level at its last scan, and links to the report and full feed. In the feed’s first full day, 54 creators sold, with their share of supply dropping by at least half since the last scan. The largest wallets also sold on 167 coins. /exits includes creator sales and locker exits. /exits all adds the largest wallets selling. It’s much busier. /exits off stops the alerts. /settings shows what your chat has turned on. Free. At most 40 posts per chat per day. Robinhood Chain only. Records, not advice. ocellus.app/exits

It's a bid slower to recover to the same developing phase once you have slowed down (was sick for some time). But the vision and the goals are the same. To show my dedication is still there got another boost, and I'm starting with the buy backs again. The developing of the best scanning/screening tool continues.

I’ve been quiet for a few days. I had some personal things to deal with, but work on Ocellus kept moving. I’m back now and picking up the pace again. Here’s what’s new: 👁️ Check my bags Paste any Robinhood Chain wallet at ocellus.app/bags to see every scanned coin it holds, its risk level, the finding that weighs most, and how its price has moved since the scan. You also get a summary of how much of the wallet sits in coins reading High or Severe. No sign-in. Nothing to connect. 🔍 Scan all Holding coins that nobody has scanned yet? One tap queues a free Quick Scan for each one. ⚡ Smarter first scans Coins people ask to scan come first. After that, the queue prioritises fresh launches, starting with those furthest along their curve. 🔧 Fixes One stuck token had blocked rescans for two days. Solana scans also paused when a data provider ran out. Both are running again, and the site now checks for both problems automatically. 📊 The track record The latest one-day snapshot covers 1,009 tokens that had at least $1,000 in them when scanned. Median price change one day after the scan: Low −13% · Moderate −31% · High −58% Share that fell 80% or more within a day: Low 18% · Moderate 11% · High 32% For the 425 tokens with a seven-day result, the median changes are still rough: Low −61% · Moderate −60% · High −76% These are memecoins. The level records what Ocellus found, not what the price will do next. Low never means safe. More coming this week. Live numbers: ocellus.app/stats Check your bags: ocellus.app/bags

@Kitxzzo It was not a planned thing ,otherwise I would have announced it.had to take care of of personal health, but this time off made me realise what was lacking

Big updates will ship in the next few days. 👁️

Thanks for flagging this. It's live now. With more than one wallet installed, the sign-in page lists them by name and you pick the one to use, so Phantom no longer loses out to MetaMask. It remembers your choice for next time. Could you try it and tell us if it opens Phantom for you?

CryptoJoe @Jo3Dang3r

a week ago

@ocellusapp Awesome work dev! Having issues signing in on the desktop, it auto-brings up metamask, when I'm using Phantom. Can you add a selector please??

1 0 1 369 0

@Jo3Dang3r Sure! The first thing on the next update will be the selector

Three things shipped on Ocellus today. 1. Every scan now shows what the price did afterwards Open any token's page. Each scan in its history has a new line, like: "After this scan: a day later the price was down 43%." We look at the price again an hour, six hours, a day and a week after every scan. Now you can see it too, good or bad, including "no market priced it any more". A risk level is only worth something if you can check it against what happened. Scans from before 6 Oct often have no line: we were missing those looks, and fixed it. 2. Phone sign-in in one tap A phone's browser has no wallet in it, so signing in meant finding ocellus.app again inside your wallet app. The Account page now has two buttons, Open in Phantom and Open in MetaMask, that take you straight there. 3. Calls from a verified wallet A call says you hold a coin and stand behind it, and the board tracks how calls did. It needed the coin in the wallet you signed in with. A wallet you've verified on your Account page now counts too. Still one call per coin, per person. Behind the scenes: Ocellus now checks itself every five minutes. Pages, the launch feed, scans, price looks, holder credits. Earlier this week we found three faults by accident. Now we get a message when something stays broken. A lot of this came from holders telling us what didn't work. Keep it coming. Full notes: ocellus.app/changelog

Paper trading accounts are still using old data. I have used the new "Patient" model for some trades, so I can personally test it in real time. ...That has been a significant win. Will be updating all paper trading accounts and the TG notifications bot soon.

Basically, the more data Ocellus collects, the better it gets at understanding which signals actually matter. That data feeds back into the system, helps refine how findings are weighted, and makes the AI more accurate over time.

Ocellus @ocellusapp

a week ago

1/7 We tested Ocellus risk levels against what happened to roughly 1,400 tokens. The result was uncomfortable, but useful: the old weighting was wrong. Tokens rated High were falling less often than tokens rated Low. Here’s what we found and what we changed. 🧵

10 2 19 599 2

@SmokeFishBass Ahhh it’s possible, that’s why the reach is down

1/7 We tested Ocellus risk levels against what happened to roughly 1,400 tokens. The result was uncomfortable, but useful: the old weighting was wrong. Tokens rated High were falling less often than tokens rated Low. Here’s what we found and what we changed. 🧵

7/7 The level’s job is to show which signals have historically been worse than others, not to say “this is safe.” That’s why we keep collecting data and testing our assumptions. The first version of a score isn’t something to defend when the results say otherwise. Full notes: ocellus.app/changelog

6/7 The most important point hasn’t changed: a risk level cannot tell you a token is safe. Around 7 in 10 tokens in this dataset had lost half their value within a week. Losses were widespread across every level. Low does not mean safe.