Multi-Bot Trading With An AIThat Will Not Hallucinate Prices.
Three simultaneous trading bots across BTC, SOL and WTI, with evidence pre-gathered before every model call. Most implementations ask the model to decide and then check reality. We reversed it: gather all evidence first, hand it to the model as structured context, and never ask it to recall market data from parametric memory.
The Actual Interface.
Language Models Do Not Know Current Prices.
That is the core problem with AI-assisted trading. A model asked what BTC is worth will answer confidently from parametric memory, and the number will be wrong.
Trend Pulse inverts the order of operations. Evidence first, decision second, so there is nothing left for the model to invent.
Evidence First, Decision Second.
Three parallel bots. BTC perpetuals, SOL perpetuals and WTI oil CFDs, all operating on Hyperliquid DEX. Each bot runs its own evidence-gathering and decision loop on an independent schedule.
Evidence pre-gathering. The model never makes a trade decision from memory. Before every call the bot pre-gathers live price data, order book depth, recent trade history, funding rates and computed technical indicator values, then passes all of it as structured context. Price hallucination is architecturally impossible.
Model routing. A routing layer sits in front of the models, so different model strategies can be tested per asset class without changing bot logic. Switching from one frontier model to another means changing a single config value, not rewriting code.
Trade assistant. A chat interface lets the operator query live position state in natural language, for example asking what the current BTC position is and why it was entered. The answer is grounded in real-time position data, so the assistant has no way to confabulate a trade that did not happen.
What The Bot Gathers Before Every Call.
Market data. Live mid price, best bid and ask, order book depth at 5 levels, 24h volume and recent trade history, fetched from the Hyperliquid API immediately before each model call.
Technicals. RSI, MACD, Bollinger Bands, ATR and EMA crossovers computed via pandas-ta on live OHLCV data. All indicator values are computed fresh and included as numeric context in the prompt.
Position state. Current position size, entry price, unrealised PnL, hourly funding rate and open order state, all included so the model reasons about actual exposure rather than hypothetical positions.
Orchestration. A Python FastAPI backend orchestrates bot loops, position tracking, Hyperliquid SDK order execution and the chat API endpoint, all as separate async tasks within one process.
- Python FastAPI
- Multi-model routing
- Hyperliquid SDK
- pandas-ta
- asyncio
Let’s Build What’s Next.
Bring the business problem. We’ll talk through what would make a difference and where to start.