Getting Started
Philosophy
The principles behind how Inserloft builds its ecosystem.
Inserloft's guiding principle is simple:
Build the future of AI infrastructure.
This means Inserloft optimizes for the ecosystem, not for any single flagship product. A few consequences follow from that:
Infrastructure over one-off models. A model is only as useful as the platform, APIs, and tooling around it. Inserloft invests in Pixel, the APIs, and the SDKs as much as it invests in training runs.
Efficiency has a place next to scale. The Kyro frontier models push capability forward, while the Baby LLM family focuses on models that are small, fast, and cheap enough to run locally. Both are treated as first-class goals, not as a "smaller version" of the same thing.
Open source accelerates adoption. Verty, Inserloft's first open-weight model, is released under the Apache License 2.0 specifically because permissive licensing encourages commercial use and wider adoption. OpenGit exists to keep publishing in that spirit.
Developers are the multiplier. Research produces models, the Platform makes them usable, and Developers turn that into products companies actually run. See the Roadmap for how these stages connect.