Changelog

You can now ask Bazzy, our new chat interface, to explain code, suggest improvements, and address concerns about changes in a Change Request.

No need to mention specific files, folders, or the entire codebase for reference—Bazzy automatically extracts the necessary context by traversing your code's AST and identifying related code elements.

Baz now integrates GPT O-1, a reasoning model designed to validate developer intent and context when detecting potential breaking changes. By analyzing the purpose and impact of changes, GPT O-1 reduces false positives, providing more accurate and meaningful insights. This enhancement ensures flagged changes are aligned with real-world scenarios, streamlining the code review process and fostering better collaboration.

With our new Slack bot integration, Baz delivers personalized updates to minimize unnecessary context switches and keep you focused. As an author, you’ll be notified on Slack when a team member comments on your change request or when it’s approved and ready to merge. As a reviewer, you’ll receive alerts when your input is needed or when one of your comments has been addressed.

No more sifting through irrelevant GitHub notifications or constantly refreshing your inbox. Baz’s Slack bot ensures you’re notified only when your action is required, helping you stay focused and efficient.

Baz now provides comprehensive detection of breaking changes in APIs

Detects critical API modifications such as removed or renamed endpoints, HTTP methods, parameters, response fields, or enum values.

Links API endpoints to code entry points, tracing parameter and return types across even in complex scenarios like destructured types or generics.

Baz now automatically reviews outdated comments and checks whether the concerns raised have been addressed by recent commits.

Reviewing large pull requests can be a daunting task, often leading to overlooked details and extended review times. The sheer volume of changes can make it difficult to form a clear understanding of the PR's impact and identify critical issues. To tackle this problem, we've introduced the Topics feature, designed to streamline and structure the review process.

Topics automatically organizes changes within a PR into meaningful groups, making it easier to navigate and evaluate the code. Each topic provides a focused view, allowing you to quickly grasp the scope of changes and their implications. This feature highlights impacted APIs, summarizes any CI checks that failed specifically for the topic's content, and identifies developers who have recently committed code to the files involved.

By breaking down complex PRs into manageable sections, Topics offers a clear mental model and structured workflow, helping you review code more effectively and with greater confidence. This enhancement ensures that even the most substantial PRs can be reviewed with clarity and efficiency, leading to faster merges and higher-quality code.

Finding the exact lines of code that cause CI check failures can be tedious and time-consuming, especially when you have to sift through extensive bash logs. To make this process more efficient, we've introduced Pinpoint Checks.

Pinpoint Checks automatically parses GitHub Actions checks and identifies the exact lines of code that triggered the failure. This eliminates the need to manually correlate failing checks with the associated code, saving you valuable time and reducing frustration.

With Pinpoint Checks, you can immediately see where things went wrong and focus on fixing the issue, streamlining your development and review process.

Code changes often impact existing APIs. With our new `Usage` analysis you can find which diff could potentially impact an existing API used by your application. This ensures reviewers are aware of the downstream changes, and can validate these changes are as intended. To view `Usage`, click on any function with an annotations label.

The analysis will consist of the definitions of the impacted APIs, invocation stats from the last 24 hours and any errors associated with this API.

Instead of continuously updating the description section of your proposed Change, we automatically generate a live descriptive walkthrough of the code changes.

Descriptions allow reviewers to instantly get up-to-speed on what’s changed since they last reviewed the code with no additional work required by authors.

We use difftastic to compare files based on their syntax, leveraging an Abstract Syntax Tree (AST) to provide a more specific diffing experience. Unlike Git diffing, Syntax diffing understands the structure and grammar of your code, ignoring benign changes and allowing developers to focus on changes with material impact.

Over 30 languages and 10 structured file types are supported.

Hi! We’re the team behind Baz, a new Source Control Management tool. We spent the last 12 months meeting developers building some of the most innovative and complex software in the world. We observed that intelligently contextualizing code is now required at every step of the development lifecycle. We think it’s time that developers collaborate using the rich context they already use to build, integrate, and ship code.