LEVERIE

LEVERIE is an AI decision-table tool that lets teams write business rules as spreadsheet-style tables and turn them into deterministic tools their LLM agents can call over MCP. Instead of burying logic in code or prompts, you define typed inputs, write the conditional rows, and publish the rule for production. It is built for developers and teams who need readable, auditable business rules for tasks like triage, review, approval, routing, and refund logic. LEVERIE provides a free web-based editor for building typed decision tables, a test environment to validate real-world scenarios, and trace inspection that shows exactly which row matched for any given input, so behavior is explainable rather than a black box. Quality-assurance checks flag coverage gaps, duplicate rows, and unreachable rows before a rule ships. Published rules are reachable through a hosted API and exposed as Model Context Protocol (MCP) tools, which is how AI agents call them without custom glue code or hand-written JSON. The result is deterministic decisions an agent can invoke with a full trace of the matched logic. A free editor is available; specific paid pricing is not published on the page reviewed.

Turning business rules into deterministic MCP tools for AI agents