rikarazome/prolog-reasoner

Overview

A tool that provides an MCP server and Python library to use SWI-Prolog as a logic calculator for large language models. It enables LLMs to perform logical reasoning by invoking Prolog queries through a standardized interface.

Best for

Best for
Developers building LLM agents that need reliable, rule-based logical inference

Use cases

• Running rule-based inference tasks within an LLM workflow
• Validating logical constraints generated by an LLM
• Combining natural language output with Prolog's deterministic deduction

How to use

Install

pip install prolog-reasoner

Tools exposed

• prolog_code
• rule_bases
• max_results
• LLM_PROVIDER
• LLM_API_KEY
• LLM_MODEL
• LLM_TEMPERATURE
• LLM_TIMEOUT_SECONDS
• SWIPL_PATH
• EXECUTION_TIMEOUT_SECONDS
• RULES_DIR
• BUNDLED_RULES_DIR
• MAX_RULE_SIZE
• MAX_RULE_PROMPT_BYTES
• LOG_LEVEL
• LLM-only
• multi-step

Tested with

Claude Desktop, Claude Code, ChatGPT

Notes

A tool that provides an MCP server and Python library to use SWI-Prolog as a logic calculator for large language models. It enables LLMs to perform logical reasoning by invoking Prolog queries through a standardized interface.

8 stars on GitHub. Last updated 2026-05-01. Licensed MIT.

Use cases

• Running rule-based inference tasks within an LLM workflow
• Validating logical constraints generated by an LLM
• Combining natural language output with Prolog’s deterministic deduction

Pros

• Leverages Prolog’s mature reasoning engine for consistent logical output
• Exposes logic calculus as an MCP server, compatible with LLM tool-use patterns
• Simple Python library for embedding Prolog query execution

Cons

• Small community and limited production usage (8 GitHub stars)
• Requires familiarity with Prolog syntax and logic programming
• Narrow focus only useful if LLM task needs formal deductive reasoning

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.