What is Open Code Review?
Open Code Review is an open-source AI code review CLI for developers who need to examine code changes or audit existing files. Open Code Review originated as Alibaba Group’s internal official AI code review assistant before becoming a community project. The review architecture combines deterministic engineering with an LLM agent rather than relying entirely on natural-language instructions.
Open Code Review reads Git diffs and sends changed files to a configurable LLM through an agent with tool-use capabilities. The output consists of structured review comments with line-level positioning. Beyond the changed lines, the agent can inspect complete files and search the repository to gather context for its findings.
What Open Code Review can do
Open Code Review supports both change-focused reviews and broader source-code audits. Deterministic components select files, bundle related changes, and match review rules to file characteristics. The agent then retrieves context and evaluates the code, while separate positioning and reflection modules refine the locations and content of review comments.
- Open Code Review reviews staged, unstaged, and untracked workspace changes.
- Open Code Review compares branches and reviews individual commits.
- Open Code Review scans entire repositories, directories, or selected files.
- Open Code Review applies rules covering null pointer exceptions, thread safety, XSS, and SQL injection.
- Open Code Review saves review results as JSON.
- Open Code Review resumes interrupted reviews and full-file scans.
How Open Code Review works
Open Code Review can be installed globally through npm, making the ocr command available in the terminal. The documentation also covers installation scripts, GitHub Release binaries, and building from source. Standard setup uses interactive provider and model configuration, including API key entry and an automatic connectivity test. Git-based operations require Git >= 2.41.
Open Code Review supports an alternative delegation workflow in which a coding agent performs the review using its own LLM. Open Code Review handles file selection and rule resolution in that mode. Documented integrations include Claude Code, Codex, Cursor, Kimi Code, OpenCode, and skill-compatible agents, while an MCP server can extend the review agent with external tools.
Who Open Code Review is for
Open Code Review serves developers reviewing local changes, teams examining branch changes, and maintainers auditing unfamiliar codebases. Full-file scanning supports repositories or directories without a meaningful diff. CI/CD integrations cover GitHub Actions, GitLab CI, GitFlic CI, and Gerrit, while the browser session viewer lets reviewers replay sessions and mark comments as fixed or ignored.
- Developers can check workspace changes before committing.
- Engineering teams can review changes between branches.
- Maintainers can audit unfamiliar source files or directories.
- Teams can incorporate automated reviews into CI/CD workflows.
Open Code Review pricing and access
Open Code Review is distributed as an open-source project under the Apache-2.0 license. Access is through the repository and documented installation methods rather than a stated product subscription plan. Standard review execution requires a configured LLM provider and model. Delegation mode instead uses the coding agent’s own LLM and requires no separate Open Code Review API key.
Open Code Review strengths and limitations
Open Code Review combines explicit file selection, related-file bundling, rule matching, and agent-driven context retrieval. The repository reports higher precision and lower token consumption than a general-purpose agent using the same underlying model, but also reports lower recall. That trade-off matters for teams prioritizing fewer false alarms: higher precision does not mean every real defect will be found.
- Open Code Review requires LLM configuration for standard reviews.
- Open Code Review requires Git >= 2.41 for Git-based operations.
- Open Code Review reports lower recall than general-purpose agents in its benchmark.
Open Code Review FAQ
What is Open Code Review?
Open Code Review is an open-source AI code review CLI that combines deterministic pipelines with an LLM agent. Open Code Review reviews Git diffs or entire files and produces structured comments with line-level positioning, using repository searches and full-file context when needed.
Is Open Code Review free?
Open Code Review is available as an open-source project under the Apache-2.0 license. Standard reviews require a configured LLM endpoint, while delegation mode uses a coding agent’s own model without a separate Open Code Review API key.
Can Open Code Review audit an entire repository?
Open Code Review can scan entire repositories, directories, or specific files through its full-file scan command. Open Code Review supports auditing unfamiliar codebases without a meaningful diff and can resume interrupted full-file scans.
Which coding agents integrate with Open Code Review?
Open Code Review documents integrations for Claude Code, Codex, Cursor, Kimi Code, OpenCode, QCA Forward, and skill-compatible agents. Open Code Review supports both reviews managed through its configured LLM and delegation to a coding agent’s own model.
How do you install and configure Open Code Review?
Open Code Review can be installed globally with npm install -g @alibaba-group/open-code-review. Open Code Review provides interactive provider and model configuration with API key entry and connectivity testing; standard reviews require that configuration unless delegation mode is used.
From the community
No reviews yet
Be the first to review this tool. What you write helps the next person choose.




