IntuitionLabs
Book a Meeting

Downloadable AI Tools for Software Engineering

Example of an AI code editor suggesting code completion (the grey text is an AI-suggested snippet, accepted by pressing Tab). Such tools can accelerate coding by predicting your next lines.

Below is a curated list of AI-powered tools that can be downloaded and run in a professional development environment (suitable for a pharma consulting firm doing AI development). Each tool supports code-related tasks like coding assistance, debugging, or refactoring. Both open-source and commercial options are included. All listed tools support multiple programming languages (including Java, Python, and TypeScript) and run on Windows, macOS, and Linux (either natively or via cross-platform editors):

1. Cursor – AI-Powered Code Editor (Freemium, Cross-Platform)

Description: Cursor is a stand-alone code editor (built on VS Code) augmented with advanced AI capabilities (Cursor AI: The AI-powered code editor changing the game). It uses cutting-edge models (GPT-4, Claude, etc.) to provide intelligent code assistance. Cursor offers multi-line code completions, natural language code editing commands, and even automatic error correction (Cursor AI: The AI-powered code editor changing the game). It effectively acts as an AI pair-programmer, helping generate entire functions and fix bugs based on plain English prompts (Cursor AI: The AI-powered code editor changing the game). This tool “works with multiple languages” and is particularly strong in JavaScript, Python, and TypeScript (Cursor AI: The AI-powered code editor changing the game), covering our required languages.

2. Codeium (Windsurf) – AI Coding Assistant (Free & On-Prem, IDE Plugin)

Description: Codeium (recently rebranded as Windsurf) is a popular AI code acceleration toolkit that integrates into your existing IDEs (Visual Studio Code, JetBrains IDEs, etc.) ( Windsurf Plugin (formerly Codeium): AI Coding Autocomplete and Chat for Python, JavaScript, TypeScript, and more - Visual Studio Marketplace ) ( Windsurf Plugin (formerly Codeium): AI Coding Autocomplete and Chat for Python, JavaScript, TypeScript, and more - Visual Studio Marketplace ). It provides context-aware code autocompletion, an in-IDE chat assistant, and other AI-powered coding features for a wide range of languages. Notably, Codeium offers a self-hosted option for enterprise, allowing use in air-gapped or on-premises environments so that no code leaves your servers (Windsurf for Enterprise | Windsurf (formerly Codeium)). This makes it suitable for organizations with strict IP or data policies.

3. Tabnine – AI Code Completion (Commercial with Local Model Option)

Description: Tabnine is an AI-powered code completion tool that plugs into many IDEs. It was an early player in AI coding assistants and supports a hybrid model approach – you can use it via cloud or opt for a completely local model for privacy (Tabnine vs Copilot: Which is A Better AI-Driven Coding Tool? | Relia Software) (Tabnine vs Copilot: Which is A Better AI-Driven Coding Tool? | Relia Software). Tabnine learns from your context and provides intelligent suggestions as you type. It supports over 50 programming languages, meeting the requirement for Java, Python, and TypeScript support (and many more) (Tabnine vs Copilot: Which is A Better AI-Driven Coding Tool? | Relia Software).

4. Sourcegraph Cody – AI Code Assistant with Codebase Context (Open Source/Enterprise)

Description: Cody is an open-source AI coding assistant by Sourcegraph that understands your entire codebase (GitHub - sourcegraph/cody: Type less, code more: Cody is an AI code assistant that uses advanced search and codebase context to help you write and fix code.). It uses embeddings and Sourcegraph’s code search to provide highly context-aware answers and code generation. You can ask Cody questions about your repository, get explanations of code, and even automate edits. It integrates with your editor and Sourcegraph’s code indexing, making it very powerful for large projects. Cody supports swapping in different AI models (OpenAI GPT-4, Anthropic Claude, etc., or even local models) depending on your needs (GitHub - sourcegraph/cody: Type less, code more: Cody is an AI code assistant that uses advanced search and codebase context to help you write and fix code.). This flexibility includes running on-premises with local LLMs, satisfying the requirement for a locally runnable LLM tool.

5. JetBrains AI Assistant – Integrated AI in JetBrains IDEs (Commercial)

Description: JetBrains AI Assistant is an AI integration available in JetBrains IDEs (IntelliJ IDEA, PyCharm, WebStorm, etc., including the new Fleet IDE) as of version 2023.3. It connects your IDE to powerful large language models (OpenAI GPT-4/GPT-3.5, JetBrains’ own models, and others) (JetBrains releases its AI Assistant and pricing – but how does it compare to Github Copilot? • DEVCLASS) to enhance your development workflow. This assistant is deeply integrated into the editor’s UI, enabling AI help within your natural coding routine.

6. Microsoft IntelliCode – AI Assistant for Visual Studio & VS Code (Free)

Description: IntelliCode is Microsoft’s AI-powered development aid integrated into Visual Studio and available as an extension for VS Code. Unlike the other tools above, IntelliCode’s models run locally on your machine once installed, which means it works even in offline or secure environments ( Visual Studio IntelliCode: AI Code Completion and Automation). It improves the existing IntelliSense autocompletion by using AI to provide smarter suggestions and can even complete a whole line of code at once. IntelliCode also assists with repetitive edits and certain refactorings in Visual Studio.

  • Key Features: In Visual Studio, IntelliCode provides whole-line autocompletions for C# (and other languages in VS Code) using an on-device deep learning model ( Visual Studio IntelliCode: AI Code Completion and Automation). It also enhances the IntelliSense lists by recommending the most likely API calls or properties at the top (based on learned patterns from thousands of open-source projects) ( Visual Studio IntelliCode: AI Code Completion and Automation). Additionally, IntelliCode can detect repetitive changes – for example, if you make a similar edit in multiple places, it offers to apply that edit throughout your code, which is a big help in refactoring scenarios ( Visual Studio IntelliCode: AI Code Completion and Automation). It also suggests quick actions for common tasks (like generating a constructor or adding a parameter) as you type ( Visual Studio IntelliCode: AI Code Completion and Automation). While not as advanced in generative capabilities as some LLM-based tools, it provides reliable AI assistance without sending code to the cloud.
  • Languages: In Visual Studio 2022, IntelliCode supports C#, C++, Java, SQL, and XAML for various features ( Visual Studio IntelliCode: AI Code Completion and Automation). In VS Code, the IntelliCode extension supports Python and TypeScript/JavaScript (and Java in preview) (Visual Studio IntelliCode) ( Visual Studio IntelliCode: AI Code Completion and Automation). This covers our key languages (Python, Java, TS) among others.
  • Platforms: Visual Studio (Windows) and VS Code (cross-platform). For VS Code, IntelliCode works on Windows, Mac, and Linux installations via the extension. (Visual Studio IDE is Windows-only, though VS for Mac had IntelliCode for C# as well.)
  • License & Pricing: Free. It’s included with Visual Studio, and the VS Code extension is free to install from Microsoft. There is no separate subscription – it’s part of the IDE tooling.
  • Download: [IntelliCode for VS Code] – available on the VS Code Marketplace (Visual Studio IntelliCode). For Visual Studio, IntelliCode is integrated; you can enable it in the VS installer or via Extensions > Manage Extensions in the IDE.

7. Void AI Editor – Open-Source AI Code Editor (VS Code Fork)

Description: Void is a fully open-source alternative to Cursor, built as a fork of VS Code with AI features integrated (Void). It was created to give developers a self-hosted, privacy-centric AI coding experience. With Void, you can use any AI model of your choice (including local LLMs) and retain full control of your data (Void). This means your code never has to leave your machine if you configure it with local models or direct API calls – Void does not funnel your code through any proprietary server.

  • Key Features: Void provides AI completions (press Tab for autocomplete similar to Copilot/Cursor) (Void), in-line natural language editing (e.g. select code and press Ctrl+K to modify via prompt) (Void), and an AI chat interface (Ctrl+L to ask questions, with the ability to reference files) (Void). Because it’s a VS Code fork, it supports all the usual extensions, themes, and workflows of VS Code. Unique to Void, you can see and edit the underlying prompts it uses, and even fine-tune its behavior (Void). It’s extensible: the community can add features or integrations – for example, plugins for code search or documentation lookup (Void). Crucially, Void lets you configure “Any LLM, Anywhere” – you can host models locally via tools like Ollama, or connect directly to external APIs (OpenAI, Anthropic, etc.) without a middleman service (Void) (Void).
  • Languages: As an editor, Void can handle any programming language (same as VS Code). The AI capabilities depend on the model you use – but popular open models like Code Llama, StarCoder, etc., support Java, Python, TypeScript and more. You have flexibility to choose models that specialize in the languages you need.
  • Platforms: Windows, macOS, Linux (since it’s based on VS Code Electron). Binaries or build instructions are provided. (At the time of writing, Void is in active development; a waitlist or invite might be required for early versions (Void).)
  • License: Open-source (the project is on GitHub (Void)). Likely MIT or similar (given VS Code’s MIT license). There’s no cost to use; community-backed development.
  • Download: [Void GitHub] – source code and releases are on GitHub. You can compile it or download pre-built releases when available. Configuration is required to link it with your chosen AI model or API. (The website voideditor.com has a “Get Access” for the latest downloads (Void).)

8. Ollama – Local LLM Runtime for Code (Open Source)

Description: Ollama isn’t an IDE or editor – it’s a command-line tool and runtime that makes it easy to run large language models (LLMs) locally on your own hardware (How to Run Open-Source LLM Models Locally | CyberArk Engineering). It’s extremely relevant for software engineering because it comes with support for several code-focused LLMs (like Code Llama, StarCoder, etc.) that you can run for code completion, code generation, or chat. Using Ollama, you can achieve ChatGPT-like or Copilot-like capabilities completely offline, which is ideal for a consulting firm needing to protect proprietary code.

Note: Popular cloud-based AI coding services like GitHub Copilot and Amazon CodeWhisperer are not listed above because they are not downloadable/self-hosted – they rely entirely on cloud APIs. The tools chosen here all have an option to run locally or on-premises, aligning with the privacy and control needs of a professional consulting environment. Each of these tools can significantly enhance software engineering tasks (from writing and completing code to debugging and documenting it) while keeping you in control of your source code.