Foundation and prototype
Authenticated projects, saved chats, multi-source research, evidence validation, AI Analytics, billing, sharing, and private-core reasoning.
PUBLIC ROADMAP
Building logical AI one measurable capability at a time.
Progress is measured by testable improvements in claim understanding, evidence use, contradiction detection, inference quality, uncertainty, and real user outcomes—not model size or demo fluency.Evaluation and expert testing
Versioned benchmarks, citation-entailment testing, contradiction accuracy, correction workflows, latency and unit-cost measurement.
Private beta and team workflows
Evidence uploads, exports, review queues, permissions, collaboration, organization controls, and controlled design-partner access.
Specialized WORE models
Permissioned, expert-reviewed datasets used for careful fine-tuning or distillation while retaining symbolic checks and independent evaluators.
Roadmap items describe current direction and are not guaranteed release dates, service commitments, or contractual promises. Priorities may change based on evaluation, safety, user needs, and technical feasibility.
BRING WORE A QUESTION THAT DESERVES MORE THAN A QUICK ANSWER.