Build Faster with Kamla

The autonomous AI coding agent that plans, executes, and refactors your code directly from the terminal.

Get Started View Commands

Why Choose Kamla?

🧠

Autonomous Planning

Give high-level instructions, and Kamla handles the breakdown, file operations, and execution.

🛡️

Multi-level Sandbox

Granular control with Read-only, Restricted, and Full access modes for maximum security.

💬

Interactive Chat

Real-time pair programming interface that understands your entire codebase context.

🚀

Provider Agnostic

Seamlessly switch between OpenAI, Anthropic, Google, and more via unified AI SDKs.

Quick Start

Get up and running with Kamla in seconds.

Installation

npm install -g kamla-cli

Initialization

Run the setup wizard to configure your preferred AI provider and API keys.

kamla init

CLI Usage

Kamla offers a powerful set of commands to streamline your workflow.

kamla chat

Start an interactive chat session for pair programming.

kamla run <task>

Execute a specific task autonomously. Example:

kamla run "Refactor src/utils.ts to use arrow functions"

kamla config

Display your current project configuration.

Global Options

Option Description
-p, --provider Set the LLM provider (openai, google, etc.)
-m, --model Specify a specific model to use
-s, --sandbox Set sandbox mode (read-only, restricted, full)

Configuration

Kamla uses a kamla.config.json file to store project-specific settings.

{
  "model": "gpt-4-turbo",
  "sandbox": "restricted",
  "approveCommands": ["npm test"],
  "blockCommands": ["rm -rf"]
}

Security & Sandboxing

Security is baked into Kamla's core. Choose the right sandbox mode for your needs.

Read-Only

The agent can only read files. Best for exploration and analysis.

Restricted

Modifications allowed, but sensitive commands require manual approval.

Full

Unrestricted access. Use only in trusted environments.