Polydictor

Prediction market analytics powered by trader performance.

Polydictor analyzes Polymarket betting markets by evaluating the track record of top token holders. Instead of predicting outcomes directly, it surfaces which outcomes are favored by historically profitable traders - giving you a data-driven edge.

How It Works

1. Track a market - Add any Polymarket event URL
2. Analyze holders - Polydictor fetches the top token holders for each outcome and scores them based on their full trading history
3. Surface signals - Key metrics are computed per outcome:
   • Prediction Rate - What percentage of a holder's past trades were profitable
   • Profit Rate - Net profit divided by total investment (ROI)
   • Weighted Profit Rate - Profit rate weighted by trade magnitude (big traders count more)
   • Total Profit - Aggregate realized PnL across all holders
4. Track over time - A continuous loop re-analyzes at a configurable interval, building a historical dataset you can chart and compare

Demo

• The prediction rate for a market outcome represents the average percentage of profitable trades among the top traders for that outcome.
• The spread is the difference between the prediction rate of two outcomes. A large spread often indicates that one outcome is more likely to win than the other.

Quick Start

Prerequisites

• Go 1.25+

Run the Web UI

# 1. Start the analysis loop (runs in background, analyzes tracked markets every 15 min)
go run .\cmd\polydictor analyze-loop -db store.db

# 2. In another terminal, start the web server (serving the API and web UI)
go run .\cmd\polydictor serve -db store.db

Open http://localhost:8081 in your browser. Paste a Polymarket URL to start tracking.

CLI Commands

serve

Starts the HTTP server and web UI on port 8081.

go run .\cmd\polydictor serve [-db store.db]

Flag	Default	Description
-db	./store.db	Path to SQLite database

API Endpoints:

Method	Path	Description
GET	/tracked	List all tracked markets
POST	/tracked?url=<polymarket_url>	Add a market to track
GET	/get-market-analysis?marketId=<id>&days=<n>	Fetch historical analysis data

analyze-loop

Continuously analyzes all tracked markets at a regular interval.

go run .\cmd\polydictor analyze-loop [-db store.db] [-frequency 15]

Flag	Default	Description
-db	./store.db	Path to SQLite database
-frequency	15	Minutes between analyses per market

analyze

Run a one-off analysis for a single market.

go run .\cmd\polydictor analyze -slug <market-slug> [-db store.db]

user

Display trading performance stats for a specific user/wallet.

go run .\cmd\polydictor user -id <wallet-address> [-db store.db]

Architecture

cmd/polydictor/       CLI entry point & subcommands
pkg/polyapi/          Polymarket API client (rate-limited)
pkg/polyflow/         Analysis engine & SQLite storage
web-ui/dist/          Single-file web dashboard (HTML/CSS/JS)

• polyapi - Wraps the Polymarket Gamma and Data APIs with per-endpoint rate limiting
• polyflow - Core analysis logic: fetches holders, scores users based on closed positions, aggregates metrics per outcome, stores results in SQLite
• Web UI - Self-contained HTML file using Chart.js for time-series visualization. No build step required

Tech Stack

Component	Technology
Language	Go
Database	SQLite
Web UI	Vanilla JS + Chart.js
Rate Limiting	golang.org/x/time
License	LGPL-2.1

AI Policy

• The backend is written and maintained by humans, and should remain that way. AI assistance is welcome, but every line of code must be fully understood and explainable by the developer.
• The web UI was largely vibe-coded using AI assistance. The prompts used during development are available in the docs directory. Each prompt used should be added to the docs directory.

License

This project is licensed under the GNU Lesser General Public License v2.1.