APEx Learns to Think in Space

When we introduced APEx in March, the argument was simple: AI should not just help people produce more stuff, it should help people think better. Not faster nonsense or prettier confusion. Actual cognitive work – the messy middle of work that sits between simple chat tasks, and coding agents.

Framing, research, synthesis, judgment, perspective – the part before execution, where the quality of the work is decide not by lines of code, but by depth of thought.

With Ape Space 2.3, we took a major step toward that vision of a super app for cognitive work.

Under the hood, APEx now runs on a completely redesigned agent loop, built specifically for large context windows, long-running cognitive tasks, and richer Whitespace reasoning. In plain English: APEx is faster, more coherent, more deliberate, and much better at staying oriented while the work gets complicated. Which, frankly, is where most interesting work lives. Research deep-dives, situational briefs, coaching sessions, strategic analysis or complex creative treatments. The kind of work where “answer the prompt” is not enough. The job is to understand what is happening, keep track of what matters, adapt as new information emerges, and steadily improve the shape of the work.

APEx now plans more clearly, narrates progress in real time, streams its work as it moves, and handles more of the cognitive plumbing on its own. Where other agent systems still expect users to tell the agent what to remember, tell it when something is context, tell it when the project changed, or when the world model needs updating – APEx detects notable background information automatically. And when you work through notes, strategies, and analysis, it can continuously refines the Whitespace Concept – the internal world model of what you are building.

That is the difference between a chat thread and a thinking environment.

We also added four new artifact types designed for serious cognitive work: Notes, Briefs, Frames, and Timelines. Notes capture freeform observations and rough thinking. Briefs turn complexity into structured situational understanding. Frames helps you – well – re-frame any problem for analysis, planning, and decision-making. Timelines map events, milestones, sequences, and change over time.

And then there is the Canvas. It allows you to arrange artifacts visually, connect them with semantic links, and build a map of how the work relates to itself – a spatial reasoning surface. APEx can read the Canvas. It can understand where artifacts sit, how they connect, what relationships exist between them, and how that spatial-semantic structure changes the meaning of the work.

Ape Space is still in beta but with v2.3 we are moving much closer to our vision of creating a co-cognitive system for people building ideas, strategies, stories, products, and futures too important to leave inside a chat window.