Dither operates Premium (50k $DITH holding requirement) and Free Demos. A demo is a demonstration of our technology with the explicit knowledge it is nascent and MVP in form. As they say in AI, “It’s the worst it is ever going to be.” We no longer have a ‘standard’ tier.
We do not offer financial advice. We only offer demonstrations of AI tools which provide information. What the end user does with this information is of their own volition.
Free Demos may be accessed from our public telegram channels.
We offer a Pump.fun bonding stream which is a collection of the most recently bonded tokens and their associated information. It is nearly anything someone could want to know about a token. We also offer a filtered stream in our premium demo suite.
API: Coming Soon
Seerlite is an on-demand iteration of our (now deprecated) original SeerBot alert system. It collects token information then gives an AI analysis of the token. We use a transformer model which operates on a 4-hour time horizon. It will tell you whether the model thinks the token is bearish, slow moving, bullish, or of unicorn status. The model is a prototype of our large time-series model, prone to mistakes, but useful as a reference for many traders. It is trained in an unsupervised fashion.
Seerlite also includes memeability ratings, name originality ratings, and description originality ratings which are built off LLM and embedding models. It combines conventional information and AI.
Otto is our mod support system in our public telegram. His job is to reduce spam, hateful content and to answer questions. Otto is an MVP so is not a fully polished product. He has shown himself useful enough we will look to productize him in the future. We may even introduce raid handling mechanics.
Premium Demos may be accessed and set through our telegram bot interface. @ditherseerbot will allow you to verify your Dither holdings without connecting your wallet!
This demo is like our pump fun bonding in the free section but uses a cluster-based filtering system to detect when tokens are outliers (not likely to rug). The anomaly cluster has a history of outperforming, but we have noticed efficiencies in the microcap market related to noise-signal which decrease this effectiveness over time. Despite this trend, we are confident in our model and our upkeep.
This is a simple alert system for dexscreener golden tickers!
The entrance/exit signals are another classification model which uses unsupervised training of a transformer-based model. It is specifically limited to larger cap tokens with a higher signal-to-noise ratio. It attempts to use the prior day of token data movement to predict whether a user should look for an entrance or an exit over the next 4 days.
This model is a precursor to our large time-series model. One of our members has claimed to have traded $800 to over $1m using the exit signals. They are also an outstanding trader in their own right, so the model was only one signal in their trading scheme.
Inflections use a volume filter and image model to read a chart. These signals are intended to show the effectiveness of Large Vision Models when applied to trading strategies. This demo illustrates how effective off-the-shelf models can be when applied to a specific system. Our AI system generates an Inflection report based on common TA methods.
Main: Return to the Main Page
Free Tier: Explore Dither Telegram and Access Free Demos
Premium Tier: Verify and Access Premium Demos
@BondedPump: Access Pump Fun Bonding Directly on Crypto Twitter
@Dither_Solana: Follow Dither AI on Twitter
Hypothesis: Check Out All Our Current Hypothesis
History: View Our Full History
AI Thesis: View Our Overall AI Thesis