Getting Started with CodeLogician

CodeLogician helps AI coding agents reason about complex software using math and logic.

Instead of relying purely on statistical reasoning from LLMs, CodeLogician introduces a logical reasoning layer that can:

  • discover edge cases
  • verify system invariants
  • explore behavioral boundaries
  • generate high-coverage test cases

Choose Your Interface

CodeLogician can be used through several interfaces depending on your workflow.

InterfaceTarget usersCode EvaluationAutoformalization AgentTUI
codelogician CLIDevelopers and agents✅ (multi-file)
VS Code extensionDevelopers (IDE workflow)
MCP serverAgent frameworks
Python libraryAdvanced automation

Recommended starting point: codelogician CLI

This lightweight shell interface lets coding agents invoke CodeLogician without installing Python dependencies.


Install the Full CLI

To install the full CodeLogician CLI:

curl -fsSL codelogician.dev/codelogician/install.sh | sh

Or install with pip (requires Python 3.12+):

pip install codelogician

Then verify the installation:

codelogician --help

Typical CodeLogician Workflows

CodeLogician supports three common workflows.

1. Direct CLI Workflow

Invoke reasoning tools directly from the command line.

Typical steps:

  • Learn the IML language using codelogician doc
  • Convert source code into formal models
  • Run reasoning with the eval command
  • Inspect proofs, counterexamples, and artifacts

2. Agent Workflow

Use the CodeLogician autoformalization agent to translate source code into analyzable models.

Examples:

codelogician agent PATH_TO_FILE

or

codelogician multiagent PATH_TO_DIRECTORY

This allows coding agents to iteratively refine models and reasoning results.


3. Server Workflow

Run CodeLogician as an interactive reasoning server.

This mode supports:

  • continuous formalization
  • multi-agent orchestration
  • live system analysis

The server can be used together with the TUI interface for interactive exploration.


What CodeLogician Provides

CodeLogician combines LLM generation with logical reasoning.

Key capabilities include:

  • Automated reasoning using the ImandraX engine
  • Formal verification of critical logic
  • Counterexample discovery when assumptions fail
  • Behavioral decomposition of complex systems
  • Test generation based on real execution boundaries

Next Steps

Continue learning how to use CodeLogician:

  • Quickstart — run CodeLogician in under a minute
  • Guides — practical workflows with AI coding agents
  • Tutorials — step-by-step walkthroughs
  • Interfaces — choose the right interface for your workflow