Skip to main content
Centaur turns high-signal trader activity into a structured, source-backed data layer for market workflows. Today, Centaur monitors curated Telegram sources and is rolling out X-backed source-message coverage where supported by the public read surface. The product vision is broader: one intelligence layer across the platforms where meaningful market signal is generated. The docs separate current coverage from that roadmap so product, API, and MCP behavior stay explicit.

​
The system

Centaur ingests monitored trader communications, classifies messages that contain trade activity, and extracts structured records from them. The output is a queryable data layer for:
  • source messages
  • trade events
  • positions
  • open-position state
  • trader and asset discovery
  • time-based performance stats
  • generated narrative summaries

​
The data

  • Structured trade events: opens, increases, decreases, and closes.
  • Current positioning: open long/short exposure by trader or asset.
  • Track records: time-based performance, win rate, asset focus, and related metrics.
  • Source transparency: source-message references, previews, and original-source links when available.
  • Generated narrative summaries: compact market narrative summaries for source windows and aggregate windows.

​
Two ways to use Centaur

Product app: discovery, profiles, Feed, Following, Assistant, alerts, API keys, and connected apps. API & MCP: read-only REST and agent-native MCP surfaces for integrations, analytics, and agent workflows. Start with Product Surfaces for the app or Quickstart for API and MCP access.
Last modified on June 24, 2026