lanter

CLI tool that uses agentic AI to convert codebases between programming languages.

Early stage — This project is under active development. Expect breaking changes.

Install

npm install -g @lanter/lanter

Getting started

1. Configure a provider

Lanter supports three AI providers: OpenAI, Anthropic, and Ollama.

Set your provider, base URL, API key, and model:

lanter config set provider ollama
lanter config set ollama.baseUrl https://ollama.com/v1/
lanter config set ollama.apiKey gw_XXXXXX
lanter config set model glm-5

3. Evaluate a codebase

Before converting, run an evaluation to assess feasibility, risks, and effort:

lanter evaluate -i ./my-project -d python

| Flag | Description | |------|-------------| | -i, --input <dir> | Source code directory (required) | | -d, --destination <language> | Target language (required) |

This produces an evaluation report with a verdict, effort estimate, risks, and blockers.

4. Run the conversion

Once satisfied with the evaluation, run the actual conversion:

lanter run -i ./my-project -o ./my-project-python -d python

| Flag | Description | |------|-------------| | -i, --input <dir> | Source code directory (required) | | -o, --output <dir> | Output directory for converted code (required) | | -d, --destination <language> | Target language (required) |

Configuration

Configuration is stored at ~/.lanter/config.json. You can manage it with the CLI:

lanter config set <key> <value>   # Set a value
lanter config get <key>           # Get a value
lanter config list                # Show all settings
lanter config model               # Interactive model picker

Config precedence (highest to lowest):

1. CLI flags (--provider, --model)
2. Environment variables (OPENAI_API_KEY, ANTHROPIC_API_KEY, OLLAMA_BASE_URL, etc.)
3. Config file (~/.lanter/config.json)

Contributing

See CONTRIBUTING.md for development setup and guidelines.

License

MIT