Skip to content

🚀 Predict Super Intelligence

Choose a tag to compare

View all tags
@arbitrageprophet arbitrageprophet released this 28 Dec 22:30
· 35 commits to main since this release
v2.0.0
508da18
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Release Date: December 28, 2025

🧠 Predict Super Intelligence

The flagship feature of v2.0.0 — a powerful multi-agent AI system that enables intelligent, team-like analysis and execution across prediction markets.

Two Modes of Operation

Mode Description Use Case
🔍 Supervised Agents analyze markets and provide recommendations. User reviews the analysis and executes via OkBet one-click links. Research, learning, manual trading
⚡ Autonomous Agents analyze markets and automatically execute trades within your budget limits ($1-$100). Automated trading, hands-off execution

Agent Pipeline

  1. Predict Agents — Deploy one or more AI agents (xAI Grok or OpenAI GPT models) to independently analyze prediction markets. Each agent can be equipped with different tools and custom commands.

  2. Predict Bookmaker Agent — When multiple agents complete their analysis, the Bookmaker acts as a "judge" that synthesizes all perspectives, weighs agreements/disagreements, and produces a consolidated recommendation with consensus metrics.

  3. Mapper Agent — Translates analysis outputs into platform-specific order parameters ready for execution.

Key Capabilities

  • Multi-Model Support — Mix and match xAI Grok (4.1, 4) and OpenAI GPT (5.2, 5.1, 4.1) models
  • Tool-Augmented Analysis — Agents can use X (Twitter) search, Web search, and Polyfactual deep research
  • Custom Commands — Direct agent focus with natural language instructions
  • Real-Time Pipeline Visualization — Watch agents work through the analysis pipeline
  • Consensus Metrics — See how agents agree or disagree on recommendations
  • Budget Controls — Set strict limits for autonomous execution

📖 For detailed architecture diagrams, see the Super Intelligence Setup Guide

🔷 DFlow Integration for Kalshi

PredictOS now supports Kalshi prediction markets through the new DFlow API integration, enabling cross-platform market analysis.

Platform Detection

The system automatically detects the platform from your market URL and routes to the appropriate data provider:

Platform Data Provider Features
Kalshi 🔷 DFlow API Event ticker resolution, Yes/No bid & ask prices, volume & 24h volume, liquidity metrics, open interest, market status
Polymarket 🌐 Dome API Market data, CLOB tokens, WebSocket feeds, order execution

How It Works

Simply paste a Kalshi or Polymarket URL into the Market Analysis terminal:

  • Kalshi: https://kalshi.com/markets/kxtechq1/... → Routes to DFlow API
  • Polymarket: https://polymarket.com/event/... → Routes to Dome API

DFlow API Features

The DFlow integration provides:

  • Event ticker and nested markets data
  • Yes/No bid & ask prices
  • Last traded price
  • Volume and 24-hour volume
  • Liquidity and open interest
  • Market status and close time

⚠️ Note: The current integration uses DFlow's development environment (dev-prediction-markets-api.dflow.net). For production endpoints and API keys, contact DFlow directly.

📦 Technical Changes

New Modules

  • supabase/functions/_shared/dflow/ — DFlow API client for Kalshi market data
    • client.ts — Base HTTP client
    • endpoints.ts — API endpoint functions
    • types.ts — TypeScript definitions for Kalshi markets

Updated Components

  • AgenticMarketAnalysis.tsx — Super Intelligence UI component
  • get-events Edge Function — Now supports both Kalshi (DFlow) and Polymarket (Dome)
  • Platform auto-detection logic for routing market URLs

🔗 Links

Built with ❤️ by the PredictionXBT team

Powered by Dome (Polymarket) & DFlow (Kalshi)

Assets 2
Loading