AI-assisted on-chain investigations. Map holders, trace funding and challenge connections. 0xbf8e5f9068ff68183b0c09bcc35b5a80aa60a430

Pinned Tweet
I don’t want to post every small update and flood the X account, so I’ll keep the bigger changes here and use the links below for everything else. Ocellus reads the chain so you don’t have to trust anyone, including us. (More chains coming soon.) Paste any Robinhood Chain or Solana token and get a report backed by the records behind it: who holds it, where the money came from, whether liquidity can be pulled, and whether the token can actually be sold. That last part isn’t based only on past trades. Ocellus simulates buying the token and selling it straight back against the current chain state, then shows what happened and what the round trip would cost. 🔎 Screener → ocellus.app/tokens 📊 Live stats → ocellus.app/stats 🗺 Roadmap → ocellus.app/roadmap 📝 Changelog → ocellus.app/changelog 🤖 Two agents testing our own rules → ocellus.app/agent
31
10
64
2,447
They’ve dominated this space for a long time, yet somehow nobody shipped this. As far as I can see, Ocellus is the first on the market with it. That’s exactly the kind of gap I want us to keep finding.
Got a picture but don’t know the token? 🔎 Ocellus lets you search by image. Choose a file, let it search the database, and explore the matching tokens. From there, open a result to look closer. For the $TripleT example, I intentionally use a different image from the coin’s icon. It still finds the match, showing how the search handles visually related images beyond exact copies. Try it with an image of your own: ocellus.app/tokens
7
3
14
278
Everything shipped on Ocellus today: ⚡ The homepage is now a live control room. Launches, scans, holder sells and paper trades appear as they happen, alongside live counters, recent findings, level performance and the six agents. 🎯 Reports now open with a one-line verdict. Risk level, the strongest finding, unread checks and a direct jump to the evidence. 🖼️ Find by picture got much smarter. Paste an image and Ocellus can now match visually similar token logos and characters, not just exact copies. 📱 Mobile got a full rework. Bottom navigation, better spacing, proper touch targets, full-screen home screen support and cards by default on Android. 📈 BTC, ETH, SOL, BNB, XRP and DOGE are now visible at the top with 24h movement, so you can see whether a move is token-specific or market-wide. 🔧 A lot of data fixes shipped too: non-ETH launch curves priced correctly, impossible holder readings rejected, sell-offs timestamped properly, and first-sale timing added. 🔗 Shared Ocellus links now have a new preview built around the live homepage. The important part is that every scan keeps adding to the same record. More wallets labelled, more tokens tracked, more history, better context for the next scan. Everything above is already live and dated here: ocellus.app/changelog
13
3
21
355
Welp, they vamped our product, and do less than 10% of what we do... Guess people will buy forced coins, instead of real working products. They can vibe code 10% of what we do and hype it, but the real data engineering and machine learning can't be.
5
1
14
304
This is live, however it does checks on more viewed tokens with higher MC. Working on a way to optimize it better, as paying X API call to review every single token, that's even 4k and is not traded is not an option. A lot of updates happened today. Will do a recap a bit later.
A new feature is coming to Ocellus: off-chain project research. The scan will start looking beyond the token itself, including things like the project’s X account, mentions, linked profiles and other public signals that can add context to what is happening on-chain. The first version is already working, but running those checks on every scan gets expensive very quickly, especially when Ocellus is processing thousands of scans and rescans. So before releasing it, I’m optimizing how the data is collected and cached so we are not paying to fetch the same information over and over again. The feature works. Now it just needs to be efficient enough to run at scale. Once that is done, it goes live.
2
2
22
499
Ocellus retweeted
I don’t want to post every small update and flood the X account, so I’ll keep the bigger changes here and use the links below for everything else. Ocellus reads the chain so you don’t have to trust anyone, including us. (More chains coming soon.) Paste any Robinhood Chain or Solana token and get a report backed by the records behind it: who holds it, where the money came from, whether liquidity can be pulled, and whether the token can actually be sold. That last part isn’t based only on past trades. Ocellus simulates buying the token and selling it straight back against the current chain state, then shows what happened and what the round trip would cost. 🔎 Screener → ocellus.app/tokens 📊 Live stats → ocellus.app/stats 🗺 Roadmap → ocellus.app/roadmap 📝 Changelog → ocellus.app/changelog 🤖 Two agents testing our own rules → ocellus.app/agent
31
10
64
2,447
Ocellus retweeted
Got a picture but don’t know the token? 🔎 Ocellus lets you search by image. Choose a file, let it search the database, and explore the matching tokens. From there, open a result to look closer. For the $TripleT example, I intentionally use a different image from the coin’s icon. It still finds the match, showing how the search handles visually related images beyond exact copies. Try it with an image of your own: ocellus.app/tokens
7
6
19
721
Ocellus retweeted
The homepage has been completely rebuilt around live statistics and data. Also the mobile version has been redesigned and polished for a more natural app feel. All feedback is welcome.
5
5
20
520
Ocellus retweeted
This was all I needed to know what the next feature needs to be. Off the roadmap, but I know it will be useful. This feature is now live. You can find the CA by just sending a picture. PS: I know you guys want videos on the updates, but I am not really good at this. So I hired someone to do new video updates. He will also help with the UX.
2
5
14
514
Today marks day 8 since I’m building this. It was the first day I was able to get 8+ hours of sleep. Now back to building.
20
4
37
721
Out of all the tools trying to copy or recreate what we’re doing, Ocellus is still one of the few built on real architecture. A real database. Real data pipelines. AI that gets better as more data comes in. We survived the first week. In a few hours, Ocellus officially hits 7 days live. And this is not something I’m building for a quick run. I want Ocellus to become the product I look back on as my life’s work. Not backing down.
23
4
30
712
After we bonded, a few copycats started popping up. Flashy dashboards, big claims, bundled launches. Some are sitting above us right now. Some have already rugged. What most of them are missing is the part that actually matters: real data engineering. Ocellus is being built to learn from the data it collects and get better over time. That means structuring it properly, understanding what is useful, and improving the signals based on what actually happens. You can generate a lot with AI, but without the engineering behind it, it usually ends up as the same polished surface with very little underneath. I’m going to keep building regardless of the market cap. Whether it is 10K or 10M does not change the reason I started this. The goal is still the same: build a product that genuinely helps people understand what they are looking at before they make a decision.
7
4
21
657
Ocellus retweeted
Just burned another 1.5% of the supply, bringing the total burned to 8%. Over the next few days, I’ll also start rolling out additional advanced paid features. A percentage of the revenue generated from those features will go back into $OCELLUS through buybacks and burns. The idea is simple: as the product grows and gets used more, part of that activity feeds back into reducing the supply.
27
9
35
727
Ocellus retweeted
Shipped today so far. 🧭 The report is rebuilt around what it actually found. Findings are no longer squeezed into a narrow column beside the chart. They now live in clear cards across the page, each marked by what it means: checked, noted, needs a look, or could not be read. That last one matters. A check that could not run should never look like a check that passed. 📊 The numbers now explain themselves visually. If a finding represents part of the supply, Ocellus shows it as a proportion instead of just throwing another percentage at you. Yes or no findings stay simple, so it is easier to understand what kind of information you are looking at without reading every line twice. ⚡ The chart is now visible as soon as the page opens. A lot of empty space was removed, especially on mobile, so the useful part of the report starts much sooner. 🐦 X account history is now part of the overview. Ocellus compares when the account was created with when the token launched, and looks at how much the account has actually been used over its lifetime. An old account with almost no activity tells a very different story from an established, active profile. 🤖 Four more paper agents are now testing individual rules. The original careful and reckless agents differ on too many rules at once, so they can show whether being selective helps, but not which rule actually mattered. The new agents each loosen exactly one rule. One allows more concentrated holders. One does not wait for a simulated sale. One buys immediately instead of waiting. One holds longer. That means when one of them takes a trade the careful agent refuses, we know exactly which rule caused the difference. If a rule never changes anything, it probably does not deserve to exist. Still paper. No real orders, no real positions.
9
9
21
481
Ocellus retweeted
I know it can become quiet in between updates. That's why I decided to create a telegram community. You can join here: t.me/ocellusportal In the mean time I boosted DEX a bit. Later today 1.1% will be burned. More updates on the website coming soon.
17
9
29
730