Open Source BSD AI Coding Tools

Tabby is a self-hosted AI coding assistant, offering an open-source and on-premises alternative to GitHub Copilot.

Claude Code is an intelligent agentic coding assistant that lives in your terminal and understands your entire codebase. It helps developers code faster by executing routine tasks, explaining complex code snippets, and managing git workflows—all via natural language commands. Claude Code integrates seamlessly into your terminal, IDE, or GitHub by tagging @claude to interact with your code context. The tool is designed to simplify development by automating repetitive work and providing instant clarifications on code behavior. User feedback and usage data are collected responsibly, with strict privacy safeguards and limited retention, ensuring no feedback is used to train generative models. Claude Code is open and actively maintained with community-driven bug reporting and feature requests. Its natural language interface makes advanced coding workflows accessible without leaving your coding environment.

VibeSDK is an open source “vibe coding” platform. VibeSDK is a project built by Cloudflare. It provides a full-stack reference implementation of an AI-driven system. Users describe the application they want in natural language, and the system generates, previews, and deploys the resulting web app. It uses Cloudflare’s infrastructure (Workers, Containers, sandboxes). It can run untrusted code safely, provide live previews, and deploy apps at scale. VibeSDK gives you the exact methodology, tools, and confidence to turn your ideas into revenue-generating products, faster than you thought possible. Live preview of generated apps in isolated sandbox containers. Support for React + TypeScript + Tailwind generation, backend routing, and modern web stack scaffolding.

Crush is a next-generation, terminal-based AI coding assistant developed by Charm, designed to seamlessly integrate with your tools, workflows, and preferred LLMs. It provides developers with an intuitive, session-based experience where multiple contexts can be managed across projects. With flexible model switching, Crush allows you to change providers mid-session while retaining conversation history. It enhances productivity by combining LSP (Language Server Protocol) support with extensible MCP (Model Context Protocol) integrations for richer coding context and external tool connectivity. Built for portability, it offers first-class support across macOS, Linux, Windows (PowerShell and WSL), and BSD systems. Backed by the Charm ecosystem, Crush is a stable, actively maintained evolution of the original OpenCode project.

Groq AppGen is an interactive web application (built with Next.js and TypeScript) that uses Groq’s LLM API to generate or modify web application code based on natural-language prompts. Essentially, you tell the app what kind of web app or page you want (in plain English), and groq-appgen will produce HTML/JSX code scaffolding, layout, and optionally application logic accordingly. It supports iterative feedback: you can refine your prompt, adjust parameters or requirements, and have the app regenerate or update the code — facilitating rapid prototyping and experimentation. For developers or non-coding designers alike, groq-appgen lowers the barrier to building full web interfaces or small apps by leveraging LLM-driven code generation rather than writing boilerplate by hand. It integrates safety/content-checking via LlamaGuard to catch undesirable outputs, and includes session management, export/share functionality, and history tracking so you can iterate on designs or revert as needed.

MiniMax-M2 is an open-weight large language model designed specifically for high-end coding and agentic workflows while staying compact and efficient. It uses a Mixture-of-Experts (MoE) architecture with 230 billion total parameters but only 10 billion activated per token, giving it the behavior of a very large model at a fraction of the runtime cost. The model is tuned for end-to-end developer flows such as multi-file edits, compile–run–fix loops, and test-validated repairs across real repositories and diverse programming languages. It is also optimized for multi-step agent tasks, planning and executing long toolchains that span shell commands, browsers, retrieval systems, and code runners. Benchmarks show that it achieves highly competitive scores on a wide range of intelligence and agent benchmarks, including SWE-Bench variants, Terminal-Bench, BrowseComp, GAIA, and several long-context reasoning suites.

Qwen2.5-Coder, developed by QwenLM, is an advanced open-source code generation model designed for developers seeking powerful and diverse coding capabilities. It includes multiple model sizes—ranging from 0.5B to 32B parameters—providing solutions for a wide array of coding needs. The model supports over 92 programming languages and offers exceptional performance in generating code, debugging, and mathematical problem-solving. Qwen2.5-Coder, with its long context length of 128K tokens, is ideal for a variety of use cases, from simple code assistants to complex programming scenarios, matching the capabilities of models like GPT-4o.

This extension is for Tabnine’s Starter (free), Pro and Enterprise SaaS users only. Tabnine Enterprise users with the self-hosted setup should use the Tabnine Enterprise extension in the VSCode Marketplace. Tabnine is an AI code assistant that makes you a better developer. Tabnine will increase your development velocity with real-time code completions, chat, and code generation in all the most popular coding languages and IDEs. Whether you call it IntelliSense, intelliCode, autocomplete, AI-assisted code completion, AI-powered code completion, AI copilot, AI code snippets, code suggestion, code prediction, code hinting, content assist, unit test generation or documentation generation, using Tabnine can massively impact your coding velocity, significantly cutting down your coding time.

CWM (Code World Model) is a 32-billion-parameter open-weights language model. It is developed by Meta for enhancing code generation and reasoning about programs. It is explicitly trained on execution traces, action-observation trajectories, and agentic interactions in controlled environments. It has been developed to better capture how code, actions, and state interact over time. The repository provides inference code, reproducibility scripts, prompt guides, and more. It has model cards, utilities, demos, and evaluation artifacts. Inference scripts and utilities for code generation tasks. Evaluation benchmarks on code, mathematics, and reasoning tasks. Demos, serving code, and evaluation pipelines.

Boost is a Laravel-first toolkit that supercharges AI-assisted development by giving assistants structured, Laravel-aware context. At its core it runs as an MCP server that exposes a battery of Laravel-specific tools, so an AI agent can explore your app, inspect code and config, and take targeted actions instead of guessing. It ships opinionated, composable guidelines tuned for popular Laravel packages, which helps keep generated code idiomatic and consistent with framework norms. The package also curates a large body of vectorized Laravel ecosystem knowledge that’s scoped to what you’ve actually installed, improving retrieval precision and response quality. It’s designed to fit naturally into existing projects, supporting current Laravel releases and modern PHP runtimes with minimal setup. Rather than trying to replace your editor or framework, Boost acts like an intelligent layer that understands Laravel’s conventions and reduces the “explain my app to the AI” friction.