Grainulator

Grainulator is an open-source evidence compiler designed to support technical decision-making by structuring claims and grading them based on the strength of their supporting evidence. The platform focuses on transforming research and technical discussions into typed statements, such as factual claims or recommendations, each assigned an evidence grade that ranges from informal assertions to results measured in production. Grainulator's compiler identifies and surfaces conflicts between claims, helping users clarify and validate the basis for decisions.

The tool facilitates evidence-tracked research sprints, allowing users to pose research questions and receive answers that are graded by the quality of their supporting evidence. It features a process where research passes are run, the compiler validates the findings, and the output consists of typed claims with explicit evidence grades and any detected conflicts. An interactive demo is available, which uses a lightweight AI model via Pollinations AI; this model does not store user data.

Grainulator is delivered as a plugin that can be installed for integration with Claude Code, supporting workflows where technical teams or researchers need to compile and assess claims systematically. The tool emphasizes privacy and simplicity, noting that it operates with zero npm dependencies. As an open-source project, it is accessible to users seeking to enhance their research and decision-making processes with transparent, evidence-based methods.

By providing a structured approach to compiling, grading, and validating claims, Grainulator addresses the need for rigor and traceability in technical decisions. Its evidence grading system and conflict detection features are tailored for environments where the quality and reliability of information are critical.