100 AI Code Generation resources for developers

A command-line chat tool that allows you to edit code in your local git repository. It uses a whole-repo map to provide context and automatically commits changes with descriptive messages.

Anthropic's agentic CLI tool designed for complex refactoring and multi-file edits. It excels at understanding project-wide dependencies and executing shell commands to validate its own changes.

An open-source IDE extension for VS Code and JetBrains that allows you to swap out LLM backends (Ollama, Anthropic, TogetherAI) and build custom autocomplete providers.

A fork of VS Code that integrates LLMs at the editor level. It features 'Composer' mode for multi-file generation and native indexing of local documentation for context-aware queries.

A terminal-based AI coding engine that handles large-scale tasks by breaking them into a plan, executing them in a protected sandbox, and allowing for diff-based reviews.

An open-source platform for software engineering agents that can interact with the browser, terminal, and editor to solve GitHub issues autonomously.

A CLI tool that coordinates AI-driven code changes by scanning your project and allowing you to select specific files to include in the LLM's context window.

An assistant that uses Sourcegraph's code graph to provide context from large, distributed codebases that don't fit into a standard local context window.

An open-source agent from Block (formerly Square) that uses a 'toolkit' approach to allow the AI to interact with Jira, GitHub, and local build tools.

An agentic tool focused on the initial scaffolding phase, asking clarifying questions to generate a full-stack application structure from a single prompt.

Use Semgrep's assistant to scan AI-generated code for common reachability patterns and vulnerabilities that standard static analysis might miss.

An automated tool that takes AI-generated code and attempts to reach 100% test coverage by generating Pytest or Jest suites based on the implementation.

Integrate Snyk into the CI/CD pipeline to specifically flag 'hallucinated' dependencies—libraries suggested by AI that do not actually exist in public registries.

A CLI tool to run test cases against your code generation prompts. Use it to ensure that model updates don't break specific coding style or logic requirements.

An automated code review tool that identifies bug risks and security flaws, providing an AI-generated suggestion to fix the issue directly in the PR.

An open-source framework for testing LLM outputs, useful for validating that AI-generated SQL or API integration code meets specific business logic constraints.

A security linter for Python that should be run as a pre-commit hook to catch common AI mistakes like hardcoded passwords or insecure 'eval()' calls.

A library to evaluate RAG-based coding assistants by measuring 'groundedness'—ensuring the AI only uses the provided documentation and not outdated training data.

A validation tool that uses JSON Schema or Pydantic models to force LLMs to return structured code metadata instead of conversational text.

A set of quality gates designed to track the percentage of AI-generated code in a repository and enforce stricter linting rules on those blocks.

Create a .cursorrules file in your project root to define specific architectural patterns, naming conventions, and libraries the AI must use.

A tool that bundles your entire codebase into a single, LLM-friendly text file, optimized with headers to help models understand the directory structure.

Use Tree-sitter to parse code into Abstract Syntax Trees (AST) before sending it to an LLM, allowing for more precise 'code chunking' in RAG systems.

A specialized text splitter that understands language-specific syntax (Python, JS, Go) to avoid breaking functions or classes across context windows.

A lightweight WASM vector search engine that can be embedded in IDE extensions to provide local, low-latency semantic search of the codebase.

A library that treats prompts as functional code, allowing you to version, test, and trace code generation prompts alongside your application logic.

A Python framework for building AI workflows that keep code generation tasks 'off-prompt' by using pre-defined tools for file I/O and shell execution.

Structure your system prompts using Markdown headers (# Instruction, ## Constraints) to improve model following for complex boilerplate generation.

Use libraries like Outlines to constrain LLM output to valid code syntax, preventing the model from generating syntax errors during high-temperature runs.

A GUI for managing the vector embeddings of your private documentation, ensuring your coding assistant has access to the latest internal API specs.