Continue

Continue is an open-source AI tool that brings quality control to the software development process by running customizable AI checks on every pull request. Teams define their engineering standards as simple markdown files, and Continue enforces them automatically as native GitHub status checks with AI-generated suggested fixes.

How It Works

Developers write checks as markdown files in a .continue/checks/ directory within their repository. Each check has a name, a description, and a prompt that instructs the AI what to look for in code changes. When a pull request is opened, Continue evaluates each check against the code diff and reports results directly as GitHub status checks.

If a check identifies a problem, it provides a suggested diff that can be accepted or rejected right from the GitHub pull request interface. This eliminates the need for generic AI review comments and focuses only on the standards teams explicitly define. There are no unsolicited opinions; checks only flag what you told them to catch.

IDE Extensions for VS Code and JetBrains

Beyond pull request checks, Continue offers IDE extensions for VS Code and JetBrains. These extensions bring AI-powered code generation, refactoring, and explanation capabilities directly into the editor. Developers can chat with their codebase, highlight code to get explanations, and generate new code using context from their entire project.

The IDE extensions support a wide range of LLM providers, including OpenAI, Anthropic, Google, and local models via Ollama or LM Studio. This flexibility lets teams choose their preferred model provider and deployment option, from SaaS APIs to fully self-hosted models running locally or in the cloud.

Source-Controlled Standards

Unlike generic AI code review tools that apply broad rules, Continue treats standards as code. Check definitions are markdown files that live alongside your source code in the repository. This means checks evolve with the codebase, get reviewed through the same pull request process, and stay version-controlled. Teams can iterate on their checks over time, making them more precise and relevant as the project matures.

Continue also provides a way to generate checks from natural language descriptions. Developers can describe what they want checked, and Continue creates the corresponding markdown check file. This lowers the barrier to establishing new standards and helps teams get started quickly.

Key Features

• Source-controlled AI checks that live in your repository alongside your code, evolving with the project
• Native GitHub status check integration with suggested diffs that can be accepted or rejected inline
• Support for any LLM provider, including self-hosted and local models through Ollama and LM Studio
• Open-source under the Apache 2.0 license with an active developer community on GitHub
• IDE extensions for VS Code and JetBrains with codebase-aware context for code generation and refactoring
• Quick generation of checks from natural language descriptions to help teams get started faster