grounded-reasoning is an open-source CLI tool designed to verify the precision of multi-hop relational reasoning in outputs from large language models and agents. It provides zero-token verification and is aimed at AI researchers and developers seeking reliable evaluation tools.
grounded-reasoning is a LLM eval & observability product. It focuses on ensuring the precision and reliability of multi-hop relational reasoning in LLM and agent outputs. grounded-reasoning is an open-source project aimed at AI researchers and developers. The project is open source (MIT). grounded-reasoning is available on the command line.
ALEXaquarius builds and maintains grounded-reasoning, and the product first shipped in 2026. The project is developed in the open on GitHub with 8 commits in the last 90 days. Among its 5 catalogued features are CLI interface, LLM verification, and multi-hop reasoning.
Latest indexed changes and source events
grounded-reasoning discovered by the PulseGate indexer
Other apps tracked under the same category.
