Intrafere Research Group Press Notes
MOTO Release News
Product announcements, release notes, and milestone updates for MOTO, the Autonomous Deep Research Harness by Intrafere.
Latest Release
MOTO 1.0.7 Autonomous Proof Generation with Lean 4 Theorem Proving
May 4, 2026
Autonomous proof discovery milestone
MOTO 1.0.7 introduces public autonomous proof generation verified in Lean 4.
Intrafere Research Group has released MOTO 1.0.7, a major step for autonomous mathematical research: MOTO now performs autonomous proof generation with the Lean 4 theorem proving language. This is the first publicly available system that solves autonomous Lean 4 proof generation with 0 user creativity required, unlike similar tools that depend on users to supply the mathematical insight, proof strategy, or intermediate creative direction.
In this workflow, the user does not need to invent the proof, choose the route through the theorem space, or manually guide the creative mathematical step. MOTO explores, proposes, validates, and reports novel proof or formulation discoveries as part of its autonomous research harness.
Lean 4 is both a programming language and an interactive theorem prover used to formalize mathematics in a machine-checkable way. Instead of asking a model to merely sound convincing, Lean 4 checks whether a proof is valid under formal logic. That makes Lean 4 important for mathematical AI research because successful formalization provides mathematical verification: the proof either type-checks against the theorem prover’s rules, or it does not.
This matters because autonomous research systems need more than fluent explanations. They need verification layers that can distinguish a mathematically valid result from a plausible hallucination. Lean 4 gives MOTO a path toward machine-checked mathematical output where proof claims can be tested directly.
Release Archive
Previous Releases
January 10, 2026
MOTO ASI Now Available
MOTO Autonomous ASI by Intrafere Research Group, is publicly available as an open source release. The system is built for high-risk, high-reward novelty-seeking mathematical research and long-running autonomous research workflows that run without user interaction after start.
The release supports multiple simultaneous models working in parallel from LM Studio, OpenRouter, or mixed local and cloud configurations. Its Aggregator and Compiler architecture enables autonomous knowledge exploration, validation, and academic-style output generation from a user prompt.
