User Guide
Last updated
Last updated
To use Exos, you must first connect a wallet.
We currently support Metamask, Coinbase wallet and any Wallet Connect compatabible wallet.
Once connected, you'll be shown the main user dashboard.
The dashboard allows you to toggle between supported networks (currently, only Ethereum is available) and shows the number of times your wallet has suffered a front-run attack.
Within each of our model-specific pages, there will be up to three sub-menus. Overview provides network-level aggregate information about the particular bot type. Explore allows you to view all addresses on the networks suspected to be a bot. Export provides an interface to download a csv file containing a list of addresses suspected to be a bot.
The primary feature of Exos is its ability to extract user-configurable lists of addresses that are suspected to be a bot.
For front-running, or sandwich attack, bots, you can configure how recently that bot has been seen operating on the network.
For high-frequency trading bots, you can configure the model's "sensitivity." The higher the sensitivity value you select, the more certain you can be that the resulting addresses are in-fact machine-controlled (i.e, bot) addresses.
Say you've built a platform that allows users to claim rewards, but you want to prevent machine-controlled addresses from interacting with your contract. You can easily configure and download this list of addresses using Exos.
In this case, you're likely to select a mid-level of sensitivity, so that you can reduce the likelihood that a legitmate user would be blocked from claiming a reward.
You can also use the dynamic plots to guide your selection. There is typically a clear "elbow point" in each curve. Beyond this point, it becomes exceedingly unlikely that an address would be misidentified by the model.
You can manually inspect the transaction histories of identified addresses in the list below.
Once you're satisfied with the configuration, simply click "Download".
If you operate a DEX, front-running bots can harm novice traders, making it harder to retain and grow a trader base. Our model identifies these addresses before they've attacked dozens or even thousands of users and Exos makes it simple to identify and block them.
Just like the HFT interface, you can configure your address list then download a CSV file.
Our unique 3D explorer allows you to inspect the bots that are active at a given time. By using machine learning techniques, we're able to create a 3D map of these addresses so that similar bots appear near one another in the space.
Once you click an address inside the visualizer, you'll find a comprehnsive list of its attacks below.
Additional transaction information can then be viewed on Etherscan by clicking the transaction hash links.
On the left-hand side, you'll find menu items for the other two main sections of the app. Each uses one of Deep3 Labs' security models to provide information and address exports from the model and the model.
For any questions or feedback on the app, reach out to us via one of our or send us an email at hello@deep3.ai.