The Camile World Model API receives the full adaptive generation

Research and product usually walk apart: what is demonstrated in the lab takes months to reach those who integrate. Version v0.10.1 shortens that path at the Camile World Model — the API service now contains, in its own base, the adaptive capabilities the model's frontier already demonstrated.

There are two capability families: understanding context before deciding how to forecast, and having specialized strategies for different data regimes. Both arrive off by default — activation is an explicit operator decision — and without touching any existing route: current integrations keep working exactly as before.

What changed

The API service now carries the same adaptive core as the model's frontier — previously restricted to the research environment.

The new capabilities can be independently and controllably activated, with each one's state auditable through a public endpoint.

The public surface did not change: no existing route, format, or behavior was altered.

What this enables

Integrators now access the same level of intelligence the CWM demonstrates in research, without waiting for a new product generation.

Teams can experiment with the capabilities in a controlled environment, measure the gain on their own data, and decide when to turn them on — at the pace of their operation.

Why it matters

The gap between what a technology can do and what its users receive is where value gets lost. v0.10.1 closes that gap for the CWM: research and API now speak the same language, with the assurance that nothing existing breaks.

In practice

In practice, a team already using the API for forecasts can enable context reading and compare, on their own scenarios, results before and after — with no migration, no rewrite, no contract surprises.

Conclusion

When frontier and product live in the same core, innovation stops being a future promise and becomes today's decision. That is the level v0.10.1 establishes.