# Introduction Synth API offers programmatic access to probabilistic price forecasts across **1-hour and 24-hour horizons** for Equities, Commodities and Cryptocurrencies, powered by a decentralized network of machine learning models. The API equips traders, developers, and institutions with predictive intelligence to gain a competitive edge, identify opportunities, and make informed decisions in dynamic markets. ## Synth API Features ### Probabilistic Forecasts, Not Point Predictions Unlike traditional forecasting APIs that return a single price prediction, Synth delivers ensemble forecasts constructed from individual models submitted by the top-performing data scientists on the network. Each model is selected based on live scoring and long-term consistency. Each model consists of 1,000 simulated price paths in each of the two forecast horizons: 1-hour (high-frequency) and 24-hour (low-frequency). The 1-hour forecast is designed for short-term trading, execution timing, and microstructure-aware strategies, while the 24-hour forecast supports broader positioning, risk management, 0DTE options. * **Full probability distributions** — Understand the complete range of possible outcomes * **Confidence intervals** — Quantify uncertainty in your predictions * **Risk-adjusted insights** — Make decisions based on probability, not guesses * **Realistic market dynamics** — Captures volatility clustering, fat tails, and mean reversion ### Powered by The World's Highest Paying Data Science Competition Synth runs on Bittensor **Subnet 50**, where 200+ machine learning models compete to generate the most accurate forecasts: * **Continuous competition** — Data scientists are scored using CRPS (Continuous Ranked Probability Score) * **Quality over quantity** — Only top-performing data scientists contribute to API responses * **No single point of failure** — Decentralized architecture ensures reliability * **Transparent performance** — All data scientists scores are publicly auditable ### Battle-Tested Accuracy All predictions are evaluated against actual outcomes using industry-standard metrics: * **CRPS scoring** — Measures both calibration and sharpness of probabilistic forecasts * **Real-time validation** — Data scientists are continuously scored against benchmark volatility metrics such as GARCH and GBM * **Leaderboard transparency** — Track top performers and their accuracy over time ## Multi-Asset Coverage The API consolidates predictions across multiple asset classes: * **Cryptocurrencies**: BTC, ETH, SOL, HYPE, XRP * **Commodities**: Gold (XAU), Oil (WTIOIL) * **Equities**: S\&P 500 (SPY), NVIDIA (NVDA), GOOGL (GOOGL), TSLA (TSLA), AAPL (AAPL) ### Meta-Model Synth miners produce forecasts and are scored by the validator on Bittensor. Then, they are ranked into two leaderboards, for the two time horizons: * **high-frequency leaderboard:** forecasts are 1h forward looking and with 1 minute increment. They are updated every 12 minutes, * **daily leaderboard:** forecasts are 24h long and with 5 minutes increment. They are updated every hour. From those 2 leaderboards, we define meta-leaderboards: aggregation of scores over a period, for better stability and considering forecast quality over time. The default meta-leaderboard are defined as: * **high-frequency meta-leaderboard:** 6-days aggregated scores * **daily meta-leaderboard:** 14-days aggregated scores On the REST API, the query parameter `days` controls the number of days to aggregate. ### Forecast Horizons All endpoints support two prediction horizons, controlled by the `horizon` query parameter: | **1-Hour** | `1h` | 60 seconds | 3,600s (1h) | Scalping, prediction markets, HFT signals | | ----------- | ----- | ------------------- | ------------- | ------------------------------------------------------------ | | **24-Hour** | `24h` | 300 seconds (5 min) | 86,400s (24h) | 0DTE options pricing, LP range optimization, risk management | If `horizon` is omitted, the API defaults to `24h`. The 1-hour horizon uses the top-performing miners from the **high-frequency meta-leaderboard**, while the 24-hour horizon uses the **daily meta-leaderboard** — each optimized for their respective timeframe. Miner real-time performance dashboard can be found here: ## Advanced Analytics & Insights Beyond raw predictions, Synth transforms probabilistic forecasts into actionable intelligence. ### Prediction Market Intelligence * Comparison with Polymarket prediction markets * Cross-market arbitrage opportunities ### Volatility Analysis * Forward-looking and realized volatility metrics * Price distribution percentiles over forecast horizon * Historical volatility context ### Options Pricing * Theoretical call and put prices derived from ensemble forecasts * Multi-strike coverage around current price ### Risk Management * Liquidation probability analysis for leveraged positions * Dynamic stop-loss levels from price distributions * Tail risk assessments ### DeFi Optimization * Optimal liquidity provider (LP) ranges for Uniswap V3 and other CLAMMs * Impermanent loss estimates * Probability of staying within LP bounds ## Use Cases ### Quantitative Trading Build sophisticated trading strategies using probability distributions: * **Market divergence detection** — Identify mis-priced contracts in prediction markets when implied probabilities diverge from Synth forecasts * **Position sizing** — Use Kelly Criterion with probabilistic forecasts * **Risk management** — Set stop-losses at confidence intervals ### Options Trading Price and trade options with theoretical fair values: * **Find mispriced options** — Compare Synth prices vs. market * **Construct spreads** — Optimize bull/bear spreads * **Hedge portfolios** — Calculate optimal hedge ratios ### Risk Management Monitor and manage portfolio risk in real-time: * **Liquidation monitoring** — Track leveraged position risk * **Portfolio VaR** — Calculate value-at-risk across assets * **Tail risk** — Understand extreme outcome probabilities ### AI & Automation Integrate with AI agents and trading bots: * **LLM integration** — Connect to Claude, GPT-4, or other AI models * **Autonomous trading** — Build AI-powered trading systems * **Natural language analysis** — Generate market commentary ### DeFi Strategies Optimize yield farming and liquidity provision: * **LP range optimization** — Set optimal Uniswap V3 ranges * **Impermanent loss forecasting** — Estimate IL before providing liquidity * **Yield comparison** — Compare expected returns across pools ## API Structure ### REST API Get started [here](/getting-started/rest-api.md). ### Websocket API Get started [here](/websocket-api.md). ### Main Endpoint Categories Prediction Percentiles — Core probabilistic forecasts * `/insights/prediction-percentiles` — Prediction percentiles (1H & 24H) Insights — Advanced analytics * `/insights/volatility` — Volatility metrics * `/insights/option-pricing` — Options prices * `/insights/liquidation` — Liquidation probabilities * `/insights/lp-bounds` — LP range optimization * `/insights/polymarket/*` — Prediction market comparisons (Polymarket) * `/insights/limitless/*` — Prediction market comparisons (Limitless) ## Support & Community * **Documentation**: You're reading it * **Discord**: [Join our community](https://discord.gg/bittensor) (Scroll down the Bittensor channels to Synth) * **GitHub**: [mode-network/synth-subnet](https://github.com/mode-network/synth-subnet) * **API Specification:** * **Email**: --- # Agent Instructions: Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://docs.synthdata.co/introduction.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.