Skip to main content
Discover what to focus on: the entities and themes worth your attention right now, surfaced by what’s actually happening in the AI race. You build a scope by reading three things Gildea has already computed, then combining them into one object. What you pick becomes the filters for Retrieve.

The recipe

Pure reads, no model calls. Each step adds one piece of the scope.
1

Read the themes

themes.list returns the theme vocabulary on two axes: value_chain (structural layers, from infrastructure up to applications) and market_force (the dynamics driving them). Each theme carries its signal count and this week versus last. Pick the themes relevant to your question. See Themes.
2

Filter the entities

entities.list filters the entity space by type, theme, scale, direction, confidence, stability, notability, and name, and sorts by sort. Filter to your themes and pick the entities you will track. direction=Rising and direction=Declining split who is gaining coverage from who is losing it. See Entities.
3

Profile your main entity

entities.get returns one entity’s full profile: its trend and reach, where it sits across the value chain, the market forces around it, and the entities it appears with most. Use it to confirm the entity is worth focusing on.
4

Combine into a scope

Put the themes, entities, and a time window into a scope object. Every other recipe takes it as input.
Install gildea, set GILDEA_API_KEY, then (pure SDK, no model calls):

What you get

A scope object: a list of themes, a list of entity IDs, and a time window. Retrieve pulls verified units across it, Refresh keeps it current, Verify traces any unit to its source, and Embed saves units beside your own data.

List themes

Get entity

Retrieve