Review Staged Changes
Analyze all your staged Git modifications with comprehensive AI-powered feedback. Detect bugs, security issues, and optimization opportunities before you commit.
Get lightning-fast, expert code reviews and AI-generated commit messages directly in VS Code. Use local Ollama models or connect to Claude, Gemini, Mistral, MiniMax, GLM, Hugging Face, or any OpenAI-compatible server.
Works with your favorite AI providers
Everything you need to improve code quality, catch bugs early, and maintain consistent commit messages across your team.
Analyze all your staged Git modifications with comprehensive AI-powered feedback. Detect bugs, security issues, and optimization opportunities before you commit.
Auto-generate conventional commit messages based on your actual code changes. Maintain consistent, descriptive commits across your entire team.
Use the lightbulb menu to explain code, generate tests, fix issues, add documentation, and get refactoring suggestions—all without leaving your editor.
Download specialized review profiles from GitHub repositories. Select multiple skills simultaneously to combine expertise—security audits, performance optimization, accessibility checks, and more.
Track token usage, generation latency, and inference costs. Monitor AI performance and optimize your workflow with detailed analytics.
Switch between local Ollama and cloud providers instantly. Use the status bar to change models on the fly without configuration hassle.
Compare and review differences between any two branches or Git references. Perfect for pull request preparation and code comparison.
Automatically excludes node_modules, lock files, and minified code. Detects formatting-only changes to reduce noise and focus on real modifications.
Override AI prompts for code reviews and commit messages with your own templates. Use variable placeholders like ${code}, ${frameworks}, and ${skills} for full control.
Review any historical commit by hash, select from recent commits, or trigger directly from the Git Graph extension context menu. Also supports commit range reviews.
Switch between General, Security, Performance, Accessibility, Educational, and Strict profiles. Compliance profiles cite OWASP, PCI-DSS, GDPR, HIPAA, SOC 2, and NIST CSF standards.
Review Pull Requests and Merge Requests directly from VS Code. Post AI-generated reviews as PR comments with summary, inline, or both styles.
Automatically review staged changes with AI before every commit. Configurable severity thresholds block commits when critical issues are found.
5-step AI pipeline: analyze diff, gather context, detect patterns, deep review, and self-critique. Goes beyond single-pass reviews for comprehensive analysis.
Run AI code reviews headlessly in GitHub Actions, GitLab CI, and any pipeline via the @ollama-code-review/cli package. Auto-post results to PRs.
Dedicated AI chat panel in the Activity Bar with full conversation history that persists across sessions. Use @-mentions to inject file, diff, and review context.
Index your codebase for semantic code retrieval. The AI retrieves the most similar code snippets and injects them as additional context for deeper, more accurate reviews.
Generate visual class, flowchart, sequence, and dependency diagrams from code changes using Mermaid.js, rendered directly in the review panel.
Track score trends, severity distributions, category breakdowns, model usage, and weekly activity. Export all data as CSV or JSON.
Encode architecture decisions (ADRs), coding patterns, and team rules in a YAML file. The AI cites specific entry IDs when flagging violations.
Automatically post review summaries to Slack, Microsoft Teams, and Discord. Configure severity-based triggers to notify your team on critical findings.
Review any file, folder, or selected text without needing Git. Perfect for legacy codebases, third-party code, or files not tracked by version control.
Copy to clipboard, save as Markdown, format as PR description, create GitHub Gists, generate commit messages, and discuss reviews in the AI chat sidebar.
Real-time, token-by-token output as the AI generates your review. Supported for Ollama, Claude, and any OpenAI-compatible server.
Use the code review functionality directly in Claude Desktop via the GitSage MCP server. 16 tools available including review, commit messages, code explanation, and more.
Get started in minutes with a simple setup process. No complex configuration required.
Search for "Ollama Code Review" in VS Code Marketplace or install directly from the command line.
Set up Ollama locally or add API keys for cloud providers like Claude, Gemini, Mistral, GLM, or Hugging Face in VS Code settings.
Use Git as normal. Stage your code changes and the extension integrates seamlessly with your workflow.
Run the review command, get AI feedback, generate commit messages, and ship better code faster.
Right-click any code selection or use the lightbulb menu to access powerful AI-driven code actions. Get instant help without breaking your flow.
Get clear explanations of complex logic and algorithms
Automatically fix bugs and apply suggested improvements
Create unit tests for your functions and methods
Generate JSDoc, docstrings, and inline comments
Choose your preferred installation method and start reviewing code in minutes.
Open VS Code, go to Extensions (Ctrl+Shift+X), and search for "Ollama Code Review". Click Install.
Open in MarketplaceUse the VS Code CLI to install the extension directly from your terminal.
code --install-extension VinhNguyen-Vincent.ollama-code-review
For local AI, install Ollama from ollama.com and pull a coding model.
ollama pull kimi-k2.5:cloud
AI Providers Supported
Major Features
Open Source
Licensed
Join developers who are catching bugs earlier, writing better commit messages, and shipping with confidence.