The company behind Jev
TypeSafe AI develops Jev, its first public System One model. It focuses on structured judgments that software can consume: selecting candidates, evaluating rubrics and judging conditions.
Jev Directory is an independent resource. For model availability, API access, billing and support, use TypeSafe’s own website and documentation.
Official TypeSafe resources
Product information, current access details and the team behind Jev.
OFFICIAL / ANNOUNCEMENTIntroducing System One Models & Jev ↗The launch article, model framing, demonstrations and evaluation caveats.
OFFICIAL / DOCUMENTATIONTypeSafe documentation ↗Current API concepts, question types and integration guidance.
OFFICIAL / LIMITATIONSJev 1.13: known limitations ↗Uneven capabilities and areas that need particular care when designing questions.
OFFICIAL / PERSPECTIVEThe TypeSafe Manifesto ↗The company’s perspective on intelligence designed for software.
OFFICIAL / CONSOLETypeSafe console ↗Account access and API credential management, subject to availability.
Community and integration resources
These resources are maintained by their respective authors and organizations. They are not all official TypeSafe properties.
A community collection of projects, integrations and discussions.
COMMUNITY / LIVE DEMOTypeSafe Typewriter ↗A writing experiment showing multiple judgments about text as it changes.
LANGCHAIN / INTEGRATIONBuilding a harness with Jev ↗LangChain’s explanation of using Jev inside an agent workflow.
INDEPENDENT AUTHOR / REFERENCESecuring a server using AI ↗Flavio Copes’s article, included as an additional AI-workflow reference.
Independent reporting
How to use these sources
Use current documentation for implementation, read limitations before automating decisions, and treat vendor demonstrations or project-reported numbers as measurements of their specific setups.
For a practical walkthrough, start with the English field guide. To find something to build with, browse the 46-project directory.