Overview

The discretionary workflow is built around four data types that work at different levels of detail. You can use them independently or layer them together depending on your process.

Live headlines → Events → Insights → Signals
(raw)            (structured)  (narrative)  (aggregated)

Each layer is more processed than the one before it. Start with signals for a fast read, drill into insights and events for context, and use live headlines when you need to stay on top of breaking news in real time.

Live headlines

Headlines are the raw feed — news articles ingested and enriched as they are published. Each one carries a sentiment score, a topic classification, and a timestamp.

Use headlines when you want to monitor breaking news for a specific asset, understand the tone of recent coverage, or see exactly what is driving a move.

Endpoint: GET /v1/headlines/feed/live/ticker/{ticker}

Events

Events are structured records of discrete market-moving occurrences extracted from the news — things like production decisions, sanctions, supply disruptions, or central bank announcements. Each event has a topic, a directional label, and a timestamp.

Use events when you want a clean view of what has actually happened, separate from the noise of general news flow. They are faster to scan than raw headlines and easier to act on.

Endpoint: GET /v1/events/feed/live/ticker/{ticker}

Insights

Insights are narrative summaries that synthesise recent headlines and events into a written briefing for a specific asset. Choose a lookback window depending on your time horizon:

Use insights when you want a fast, high-quality briefing without reading through dozens of individual headlines yourself.

Endpoint: GET /v1/insights/analyst/{ticker}?analyst_type=weekly

Signals

Signals aggregate sentiment and event data into a single directional view for a ticker — bullish, bearish, or neutral. They are the most compressed form of the data: one number that reflects where the intelligence is pointing right now.

Use signals as a first filter to scan across multiple assets and identify which ones have the most notable current dynamics.

Endpoint: GET /v1/signals?tickers={ticker}

Forecasts

Forecasts are forward-looking model outputs for a ticker, generated from the underlying sentiment and event data.

Endpoint: GET /v1/forecasts/{ticker}

Next steps

• Daily Workflow — simple patterns for using these data types in practice
• API Reference — full parameter details for all discretionary endpoints
• Recipes — end-to-end worked examples