GitHub Copilot
Code completion, chat and agent mode in VS Code, JetBrains IDEs and more, plus a coding agent that works from GitHub issues.
- Free plan or free version
- Works in VS Code
- Works in JetBrains IDEs
IDE assistants add AI to the editor you already use. They install as an extension in VS Code, JetBrains IDEs, Visual Studio or other editors and give you inline completion, a chat panel that can see your open files, and quick actions such as explaining a function or writing a test. Many now include an agent mode as well, so the line between an assistant and an AI editor is getting thinner.
For a team, the main advantage is that nobody has to change tools. The editor, debugger, plugins and settings stay the same, and the assistant is one more panel. That also makes it easier to compare a few assistants on the same codebase, since switching is a matter of enabling a different extension.
The differences that matter most show up in three places. Context is the first: some assistants only see the open file, while others index the whole repository or connect to docs and tickets. Model choice is the second: some lock you to their models, while open tools such as Continue and Cline let you bring hosted or local models. Deployment is the third: a few, such as Tabnine, can run inside your own cloud or data center for teams that cannot send code to a vendor.
What a featured listing costs ยท from $99 a month
Code completion, chat and agent mode in VS Code, JetBrains IDEs and more, plus a coding agent that works from GitHub issues.
Code assistant for IDEs focused on privacy, with options to run in your own cloud or on your own servers.
AWS's coding assistant for IDEs and the command line, with chat, code suggestions and help with AWS services and upgrades.
Open-source assistant for VS Code and JetBrains that lets you plug in the models you choose, including local ones.
Open-source VS Code extension that acts as a coding agent, asking for your approval before it edits files or runs commands.
Coding assistant and agent aimed at large codebases, with chat and completions in VS Code and JetBrains IDEs.
Before you pick one, try it on code your team actually maintains, in the languages you actually use. Check the admin controls too: seat management, policy settings, audit logs and whether suggestions that match public code can be filtered out.
It also helps to separate completion from chat when you evaluate. Completion is judged in the flow of typing: does it guess the next few lines correctly, and does it get out of the way when it does not? Chat and agent features are judged on bigger tasks: explaining a module, writing a test for a tricky function, or making a change across files. A tool can be strong at one and average at the other.
Plan the rollout like any other developer tool. Start with a small group, collect examples of where the assistant helped and where it got in the way, and write a short team guide covering what not to paste, how to review suggestions and which settings to turn on. That guide does more for quality than the choice of vendor.
GitHub Copilot, Tabnine, Amazon Q Developer, Continue and Augment Code all support both. Cline is focused on VS Code. Check each tool page for the editors it lists, since support for smaller editors varies.
Some can. Continue and Cline let you point the assistant at a local model served by a tool such as Ollama. Hosted assistants generally use the vendor's own models, sometimes with a choice between several.
Not to start. Several have free plans with limits, and open-source extensions are free to install, though you may pay a model provider for the API calls they make.
Data handling and retention, whether code can be used for training, admin and policy controls, support for single sign-on, and whether a self-hosted or private cloud option exists if your code cannot leave your network.