Lynecode: AI coding agent, built for the terminal

Lynecode is an AI coding assistant from LyneLabs. It brings a fast terminal UI, built-in project indexing, an opinionated toolset, and optional AI models (OpenAI, Gemini, Azure OpenAI) with your keys. Run it on any folder and work from a powerful chat workflow.

• Integrated version control safety net: Lyne maintains its own backup snapshots (time_machine in your target path). You can explore diffs and restore changes without relying solely on git.
• Built-in analyzers and guards: Bug and code quality analyzers help spot issues quickly (without exposing internal implementation). Optional static checks are available via extras.
• Context-rich workflow: Attach files/folders inline, fuzzy-find by name, and navigate conversations with a clear menu.
• Model agnostic: Bring your own keys for OpenAI, Google Gemini, or Azure OpenAI and switch models on the fly.

Python 3.9+ is recommended.

• Run on a specific path:

• Or from your current working directory (Lynecode resolves to an absolute path):

Once launched, you’ll see a banner and a prompt. Type your request or use the in-app menus:

• Show navigation menu: /menu
• Help: /help
• Switch conversations: /chats
• Model settings: /model or /switch <model>
• Quit: quit

• Attach a file: add /file:<name_or_term> inside your message
• Attach a folder: add /folder:<name_or_term> inside your message Lynecode will fuzzy-match and help you pick the right target, then keep attachments visible after responses.

• Open the model menu from navigation (/menu → “Model Settings”) or run /model.
• Switch directly with /switch <model_name>.
• The default model is used if configured.

Supported providers (bring your own keys):

• OpenAI (e.g., gpt-4.1, gpt-5)
• Google Gemini (e.g., gemini-2.5-flash, gemini-2.5-pro, with optional “-thinking” variants)
• Azure OpenAI (your deployment name)
• OpenRouter (e.g., x-ai/grok-4-fast:free, openai/gpt-4o via OpenRouter)

Launch the API Configuration from the navigation menu:

• /menu → “API Configuration”

Follow prompts to set keys:

• OpenAI: API key (sk-...)
• Gemini: API key
• Azure OpenAI: API key, endpoint, API version (optional), and deployment name

Keys are stored securely via Lynecode’s config manager and masked in the UI.

Note: After updating any API key or provider configuration, restart Lynecode to apply changes.

Lynecode automatically creates a hidden time_machine folder inside the target directory. It’s your safety-net:

• Auto snapshots around operations
• View diffs and restore via the Version Control menu
• Works even if the target folder is not a git repo

Open from navigation:

• /menu → “Version Control”

• Intelligent file edits (create/replace/delete blocks)
• Project-wide searching (grep/AST-backed and fuzzy indexing)
• Web reading and search helpers
• Conversation history with attachments
• Rich terminal output where supported; plain fallback otherwise

• Safety first: Lynecode’s built-in snapshotting means quick rollback without external VCS.
• Deep project tooling: File block edits, AST/grep, and fuzzy search integrated in one place.
• Menu-driven UX: Rich menus for models, API setup, attachments, version control.
• Provider-agnostic: Switch between OpenAI, Gemini, or Azure OpenAI at runtime.

• If terminal colors aren’t showing, Lynecode falls back to plain text automatically.
• Ensure your API keys are valid (placeholders are detected and blocked with helpful messages).