Your new AI-powered coding assistant that can write you code, answer your questions, and level up your productivity. Available for free today in our VSCode, JetBrains, Eclipse, Visual Studio, and XCode extensions.

Generate

Codeium Chat can generate whole functions and applications for you. It's easy, fast, and free.

Explain

Diving into a foreign codebase? With the push of a button our assistant will explain everything you need to know.

Debug

Need to fix a bug or add a new feature?Or want to refactor or even translate your existing code? Make it prettier?Chatwith our assistant and ask for what you want.

Ask

Easily ask questions about your frameworks, libraries, APIs and overall tech stack, all from the comfort of your editor.

An AI Assistant for All Workflows

Full Stack Engineer

Make this component strongly-typed and use hooks for state management

Backend Engineer

Write me a unit test for when a request has empty required fields

Student

Explain multi-threading and how it's used here in this Go snippet

Software Engineer

Add docstrings for all public methods in this class

Game Developer

Write a function to get the distance between character and checkpoint

Frontend Developer

Make this button wider and glow on hover

Software Engineer

How do I undo a commit in the Git CLI?

Web Developer

Where do we set our CORS headers? What domains are whitelisted?

Software Engineer

What's the command to run my pre-commit hooks before committing?

Infrastructure Engineer

Bundle these individual data requests into a single BigQuery upload

AI suggested.

Human accepted.

We're confident in our AI-generated code, but you are still in the drivers seat. Codeium offers an intuitive way to accept, reject, or modify what our models generate.

Our Chat system is the best AI Chat on the market

The best models, the best reasoning engine, and the best UX, all in your IDE of choice so that you can be the best dev possible.

Leading model optionality

Family of in-house models, open-source based models, and third-party API-based models. Get the best results.

Optimized for coding tasks

Crafted to excel with full and multi repository context awareness and backed by our state-of-the-art reasoning engine. Get grounded results.

Broadest availability

Fully functional across the widest range of IDEs. Get results where you want.

Better than ChatGPT at code-related questions

See how Codeium Chat stacks up to the competition

	Codeium	Copilot	Amazon Q
Availability
Visual Studio Code	√	√	√
JetBrains	√	√	√
Visual Studio	√	√	√
Eclipse	√	×	×
XCode	√	×	×
Models
Fast, code-biased models	√	×	√
GPT-4 + Claude 3.5 Sonnet	√	√	×
Llama 3.1 405B based code model	√	×	×
Context * = feature only available if your code is stored in GitHub SaaS
Full Repository Context Awareness	√	√ *	×
Multirepository Context Awareness	√	×	×
@-mentions	√	√ *	×
Context Pinning	√	×	×
Custom Chat Context	√	×	×
Inline Citations	√	×	×
Features
Code Lenses	√	×	×
Highlight + Right Click for Shortcuts	√	√	√
Conversation History	√	√	×
Slash Commands	√	√	√
Chat Insights	√	×	×

Codeium Live（new）

Chat with the most up-to-date version of your tech stack, more efficiently and accurately than any other tool.

Your favorite libraries, free, forever up-to-date right in your browser

Codeium, Bun, LangChain, NextJS, Numpy, Prometheus, Pytorch, React, Sentry, Svelte, SvelteKit, Threlte, Whisper, and many more coming soon.

Don't see your favorite library here? Let us know on Discord

The @ superpower

Avoid copy-pasting code and help guide the AI to what you know to be relevant by referencing your functions, classes, files, directories, and repositories. Simply type “@” in the chat input box and refer to these entities by name.

Context Pinning

Want all results in Chat to refer to some central libraries, best practices, or other relevant context? Just pin the function, class, file, directory, and/or repository, and Codeium will know to pay extra attention to them when responding.

Inline Citations（new）

Codeium's reasoning engine helps pull in the most relevant context from across your codebase to ground responses. With inline citations, you can trust the responses by clicking into inline links to see where Codeium is pulling information from.

More Features, Shortcuts, and Insights（new）

It's the small things that add up. Custom chat context, chat history, slash commands, inline context pinning, chat insights, copy to clipboard, and a whole lot more.

Frequently Asked Questions

How does Codeium Chat work?

Codeium Chat seamlessly integrates the powers of open-ended conversation with IDE context. Besides allowing familiar interactions like those with ChatGPT, users can use smart suggestions to perform common actions such as adding documentation to functions or refactoring code. Under the hood, Codeium Chat has a variety of models to choose from. There is our Base Model (Llama 3.1 70B based, fine-tuned in-house), Codeium Premier (Llama 3.1 405B based, fine-tuned in-house), as well as OpenAI's GPT-4o and Anthropic's Claude 3.5 Sonnet. For self-hosted enterprise customers, we are able to provide Chat via our own Chat models, as well as provide optionality to integrate with private endpoints of leading model API providers.

Who should use this?

Codeium does not replace the software engineer, leaving the developer in charge and responsible for any code generated. Codeium does not test the code automatically, so a developer should carefully test and review all code generated by Codeium. So while anyone can use Codeium, we recommend it especially for people who already have fundamental knowledge of software engineering and coding. It's never great to be dependent on anything, even superpowers.

What IDEs and Languages have chat?

Codeium Chat is currently only on VSCode, JetBrains IDEs, Visual Studio, Eclipse, and XCode, but we will be rapidly supporting more IDEs in the near future. Codeium Chat will work on any language, but the CodeLens suggestions above functions are available for only common languages, which includes Python, JavaScript, TypeScript, Java, Go, PHP, and more.

√

Quick Navigation;