ThinkReview

Hey Indie Hackers! We just shipped ThinkReview Cloud 2.2.0, adding search grounding to our AI code reviews for GitHub, GitLab, Azure DevOps, and Bitbucket.

Frontier models, libraries, CVEs, and API IDs ship on a weekly cadence now. A reviewer that cannot look anything up will sound completely sure while being weeks behind. Previously, a review might flag a brand-new fallback chain as using "nonexistent" model IDs simply because those models shipped in the weeks before the review, resulting in confident, stale feedback.

With search grounding in Cloud 2.2.0, when a finding depends on what exists today—like a new model name, a package version, or a deprecation—the review can check current vendor docs and live sources instead of inventing a story from memory. It combines full-context repo awareness with live web facts.

Nothing changes in how you run a review: open a PR/MR on your platform of choice, pick a cloud model, and search grounding handles the rest automatically (local reviews remain strictly local). Check out the release details at https://thinkreview.dev/blog/2026-08-23-search-grounding-thinkreview-cloud. I'd love to hear your thoughts or feedback on handling stale AI context in your code review workflows!

Up until now, Edge users could run ThinkReview by installing it from the Chrome Web Store and allowing extensions from other stores. While that worked, many enterprise teams and Windows developers standardized on Edge face strict IT policies regarding third-party store flags.

Now, you can install ThinkReview directly with a native 'Get' button from Microsoft's official store: https://microsoftedge.microsoft.com/addons/detail/thinkreview-ai-code-revi/hbcanlmjippkplohigflbknfcajdkglk

### What is ThinkReview?

It’s an AI code review copilot that runs directly in your browser side panel on pull requests and merge requests for GitHub, GitLab, Azure DevOps, and Bitbucket—no CI pipeline setup or admin privileges required. It supports custom review agents, full repository context, chat on PRs, and lets you use cloud frontier models, local Ollama, or OpenRouter.

### Why this matters

For developer tools, official store listings make adoption smooth. If your company standardizes on Edge and requires Microsoft-vetted store listings, this removes friction for getting AI code reviews into daily workflows.

If you use Edge, give the new listing a spin and let us know how it works for your team! You can drop feedback or check out the open-source repo at https://github.com/Thinkode/thinkreview-browser-extension.

Thanks!

Hey Indie Hackers! We just shipped a major milestone for ThinkReview, our AI code review tool for GitHub, GitLab, Azure DevOps, and Bitbucket. ThinkReview MCP (v1.3.0) is now live on the Official MCP Registry at registry.modelcontextprotocol.io under the namespace io.github.Thinkode/thinkreview!

What we shipped:

We've pointed the registry record at our hosted Streamable HTTP endpoint (`https://mcp.thinkreview.dev/v1`). This makes it vastly easier for developers to discover our server without hunting through GitHub repos or documentation.

With ThinkReview MCP, users get:

- Live PR/MR reviews via the review_url_code tool supporting GitHub, GitLab, Azure DevOps, and Bitbucket URLs.

- Secure OAuth (recommended) or portal Bearer auth.

- Native compatibility with Cursor, Claude Code, Claude Desktop, and GitHub Copilot.

- The exact same review engine and credits as our browser extension.

Why it matters:

Getting listed on the official registry bridges the gap for AI-native developers who live inside tools like Cursor and Claude. Instead of forcing manual local npm installations, our remote MCP architecture lets users configure a simple endpoint, authenticate, and instantly paste a pull request URL to get a comprehensive code review.

If you're building or using MCPs, we also wrote a transparent breakdown of how we published a remote MCP without requiring an npm package: https://thinkreview.dev/blog/how-i-listed-my-mcp-on-the-official-mcp-registry

You can check out the listing and get started via https://thinkreview.dev/blog/2026-08-15-thinkreview-mcp-now-on-official-mcp-registry. I'd love to hear your thoughts or answer any questions about our remote MCP setup and auth flow!

I built a browser extension that adds AI code reviews directly to Azure DevOps Pull Request pages. No copying code, no switching tabs.

The problem: I was constantly switching between Azure DevOps and external AI tools for code reviews. Copying code blocks, losing context, breaking flow.

The solution: An extension that adds an "AI Review" button directly on PR pages. It uses Azure DevOps Git APIs to fetch actual code diffs and provides instant reviews right where you're working. Works with both dev.azure.com and visualstudio.com URL formats.

Key features:

- Zero context switching - everything happens in-browser

- Persists during SPA navigation

- Uses Git APIs for accurate code analysis

- Simple setup (just add PAT token)

Demo: https://thinkreview.dev/media/thinkreview-azure-devops-full.gif

I've been using it for a few weeks and it's saved me significant time. Would love feedback from other Azure DevOps users:

- Does this solve a real problem for you?

- What features would make it more useful?

- Any Azure DevOps-specific quirks I should be aware of?

Available: https://chromewebstore.google.com/detail/thinkreview-ai-code-revi/pfmgefgllkkjbkfdnkccjfajkkjdchkm

More details: https://thinkreview.dev/blog/i-built-an-ai-code-review-tool-for-azure-devops-that-eliminates-context-switching

• What makes ThinkReview unique?

  • Its biggest differentiator is the zero-configuration setup. Unlike many competitors, ThinkReview works instantly as a browser extension without needing complex CI/CD integration, setup tokens, or granting access to your entire GitLab/GitHub project. It enhances the developer's existing workflow directly in the browser.

• Why should a person choose ThinkReview over its competitors?

  • You should choose ThinkReview for its simplicity and speed. It's the fastest way to start with AI code reviews—install the extension, and you're done. It empowers individual developers to improve their reviews immediately, without the overhead and enterprise-level pricing of deeply integrated tools.

• How would you describe your primary audience?

  • Our primary audience consists of individual developers and small to medium-sized teams using GitLab and GitHub. It's perfect for developers and team leads who want to improve code quality and efficiency without a complex, top-down implementation process.

• What's the story behind ThinkReview?

  • The story is simple: it was born from the founder's frustration with code reviews being a slow, repetitive bottleneck. The goal was to create a tool that was easy to install, worked out of the box, and acted as a true AI copilot to assist the human reviewer, not just an automated bot.

• Which are the primary technologies used for building your product?

  • The core of the product is a browser extension built with JavaScript, HTML, and CSS. It communicates securely with major AI and Large Language Model (LLM) APIs to provide its intelligent feedback.

• Who are some of the biggest customers of ThinkReview?

  • As a recently launched tool, our primary user base consists of individual developers and growing teams who have adopted ThinkReview to streamline their personal and team workflows. We are focused on building this grassroots support before targeting large enterprise customers.

• Proactive AI with Suggested Comments & Smart Questions: ThinkReview now goes beyond just responding. It proactively generates professional code review comments and suggests insightful questions to help you and your team understand merge requests faster and more deeply.

• Speak Your Language (For Good!): This was a big one. You can now set a persistent language in the settings. Set it once, and all AI feedback will be in your chosen language automatically. No more repeating "Czy możesz odpowiedzieć po polsku?"!

• Expanded Browser Support: We're now officially on Microsoft Edge and all other Chromium browsers (Opera, Brave, etc.), making it accessible to more teams.

• A Faster & Smarter AI Core: We've upgraded the underlying model, making it significantly faster and more intelligent than ever before.