Continue

What is Continue?

Continue is a software development tool designed to automate and standardize code review processes. It integrates directly with GitHub to run AI-driven checks on every pull request. The core concept involves defining specific code quality and security standards as simple markdown files, which are stored in your repository. These checks are then executed automatically by an AI agent against the code changes in a pull request. The results appear as familiar GitHub status checks, passing if the code meets the defined criteria or failing with a suggested fix if it does not. This approach shifts code review from a purely human, opinion-based process to a more consistent, automated system that catches common, describable issues, allowing human reviewers to focus on higher-level architectural and design decisions.

Main Features

1. Source-Controlled Checks: Define standards as markdown files within your repository's .continue/checks/ directory, making them version-controlled, reviewable, and owned by the team.

2. Automated PR Analysis: Runs as a full AI agent on every pull request, analyzing the diff to apply your defined checks.

3. Focused Enforcement: Designed to catch only the specific issues you define, avoiding unsolicited opinions or generic feedback for more reliable and predictable results.

4. Integrated Fix Suggestions: When a check fails, it provides a concrete code fix that developers can accept or reject directly from the GitHub interface.

5. Mission Control Dashboard: Offers a central dashboard for managing checks, viewing metrics, and monitoring performance across projects.

Use Cases

1. Automated Security Reviews: Enforce checks for hardcoded secrets, missing input validation on new API endpoints, or unsafe SQL query construction.

2. Code Consistency Enforcement: Ensure naming conventions, documentation standards, or specific architectural patterns are followed across all contributions.

3. Mechanical Review Automation: Free senior engineers from repetitive first-pass review tasks, allowing them to focus on complex logic and design judgments.

4. Pre-Merge Quality Gates: Implement automated gates that must pass before code can be merged, reducing bug introduction and maintaining codebase health.

5. Team Onboarding: Provide clear, automated standards for new team members, helping them adhere to team practices from their first pull request.

Supported Languages

1. The tool itself and its check definition system are language-agnostic. Checks are written in plain English within markdown files.

2. The underlying AI agent can analyze code written in virtually any programming language, as the analysis is based on the semantic understanding of the code changes provided in the pull request diff.

Pricing Plans

1. Starter Plan: Priced at $3 per million tokens (input and output). This is a pay-as-you-go plan for creating and running AI agents, buying credits for frontier models, and connecting integrations.

2. Team Plan: Priced at $20 per seat per month, which includes $10 in credits per seat. It offers all Starter features plus centralized management for private agents, team controls, and Gmail/GitHub SSO login.

3. Company Plan: Custom pricing for enterprises. Includes all Team features plus custom SSO (SAML/OIDC), bring your own API keys (BYOK), commitment, invoicing, and SLA.

Frequently Asked Questions

1. Q: What is Continue?

A: Continue runs AI checks on every pull request. Each check is a markdown file in your repo that shows up as a GitHub status check — green if the code looks good, red with a suggested fix if not.

2. Q: How does it work?

A: You define checks as markdown files in .continue/checks/. Each file has a name, description, and a prompt telling the AI what to look for. When a PR is opened, Continue runs each check and reports the result.

3. Q: What can I use checks for?

A: Checks can be used to flag security issues (like hardcoded secrets), enforce code style, ensure documentation is present, or any other standard describable in a prompt.

4. Q: Where do the checks run?

A: Checks can be run locally via the CLI, directly on pull requests via the continue.dev/check interface, and integrated into your CI/CD pipeline.

Pros and Cons

Pros:

1. Provides consistent, automated enforcement of team-defined coding standards.

2. Integrates seamlessly into the existing GitHub workflow with familiar status checks.

3. Saves significant developer time by automating mechanical review tasks.

4. Standards are version-controlled and transparent, living alongside the code.

5. Offers focused feedback, reducing noise compared to broad AI code review tools.

Cons:

1. Requires upfront effort to define and tune effective checks for your team's needs.

2. Effectiveness is dependent on the quality and specificity of the prompts written for each check.

3. Primarily integrates with GitHub; teams using other version control platforms may have limited support.

Recommendation Rating

8/10 (A powerful tool for teams seeki