Purple Flea for MindsDB

How MindsDB + Purple Flea Works

MindsDB is an AI-powered database layer that understands natural language, runs machine learning models, and connects to external services through its engine abstraction. Its HTTP engine lets you map any REST API as a virtual database table — meaning you can SELECT from and INSERT into Purple Flea as if it were a relational database.

The combination is powerful: MindsDB's LLM functions can analyze market sentiment from text, predict price movements from historical data, and immediately pass those signals to Purple Flea's trading and casino APIs — all within a single SQL workflow.

SQL-Driven Trading with MindsDB

Below is a complete MindsDB SQL workflow: create the Purple Flea HTTP integration, query live market prices, run an LLM sentiment analysis, and execute a trade — all in SQL. No Python orchestration needed.

-- Step 1: Register Purple Flea as an HTTP database in MindsDB
CREATE DATABASE purple_flea_api
WITH ENGINE = 'http',
PARAMETERS = {
  "base_url": "https://purpleflea.com/api/v1",
  "headers": {"Authorization": "Bearer your-api-key"}
};

-- Step 2: Query live market price
SELECT symbol, price, change_24h, volume_24h
FROM purple_flea_api.markets
WHERE symbol = 'BTC-PERP';

-- Step 3: Create an LLM model for sentiment analysis
CREATE MODEL market_sentiment
PREDICT sentiment_score
USING
  engine = 'openai',
  model_name = 'gpt-4',
  prompt_template = 'Rate the bullishness of this crypto market data
    on a scale from -1.0 (very bearish) to 1.0 (very bullish).
    Data: {{market_data}}
    Return only the numeric score.';

-- Step 4: Join live market data with AI sentiment
SELECT
  m.symbol,
  m.price,
  m.change_24h,
  s.sentiment_score
FROM purple_flea_api.markets AS m
JOIN market_sentiment AS s
WHERE m.symbol = 'BTC-PERP';

-- Step 5: Execute trade when sentiment is strongly bullish
-- (Run via MindsDB job scheduler or trigger)
INSERT INTO purple_flea_api.trade (symbol, side, size_usd, leverage)
SELECT
  'BTC-PERP',
  'long',
  100,
  5
WHERE (SELECT sentiment_score FROM market_sentiment
        WHERE market_data = 'BTC 24h change: +4.2%, volume spike: 180%'
       ) > 0.7;

-- Step 6: Check wallet balance
SELECT token, balance, usd_value
FROM purple_flea_api.wallet
WHERE chain = 'ethereum';

-- Step 7: Create an automated job that runs every 15 minutes
CREATE JOB btc_sentiment_trader
START '2026-03-07'
EVERY 15 minutes
DO (
  INSERT INTO purple_flea_api.trade (symbol, side, size_usd)
  SELECT 'BTC-PERP', 'long', 50
  FROM market_sentiment
  WHERE market_data = (SELECT CONCAT(symbol, ' price: ', price) FROM purple_flea_api.markets WHERE symbol = 'BTC-PERP')
    AND sentiment_score > 0.65
);

AI + Financial Actions

MindsDB's AI layer connects seamlessly to Purple Flea's execution layer. Each prediction output maps to a specific financial action — no glue code, no separate orchestration service.

Available Financial APIs

Purple Flea exposes six production services as HTTP endpoints. Each maps to a MindsDB virtual table — SELECT to read state, INSERT to execute actions.

Use Cases

The MindsDB + Purple Flea combination unlocks workflows that are impossible with standalone tools. Your data warehouse becomes a live trading system.

Related Integrations