Urgent.News

What's breaking now, across thousands of outlets.

Tech

Meet LiveReview: Auto Approve 30-50% of Your Low-Risk PRs, Safely

Earlier this year, we built a small tool called git-lrc . It fixed one problem: engineers committing AI-written code they had not read. git-lrc put a review in front of every commit, right in the terminal. In February, it reached #3 Product of the Day on Product Hunt, and it now has 1.5k+ GitHub stars. As teams adopted it, a bigger problem showed up one step later, at the pull request. AI made…

LiveReview is a new code review system designed to automatically approve 30-50% of low-risk pull requests while ensuring risky changes receive appropriate attention. The system analyzes code changes using three methods: semantic LLM analysis, custom ML models trained on actual production incidents, and static codebase analysis. LiveReview assigns a risk score to each change, allowing the most critical 30-50% to be approved automatically.

For the remaining changes, LiveReview ranks them based on risk, ensuring senior engineers focus on the highest-risk code first. The system also provides checkpoints at various stages, such as commit, push, PR, CI/CD, and scheduled checks, allowing teams to customize which stages they want to enforce. Additionally, LiveReview offers a CI/CD gate feature that allows teams to write custom conditions to block or allow changes based on the review findings.

LiveReview is self-hostable and integrates with various git providers, as well as multiple AI providers, making it a flexible solution for any engineering team. By automating the review process and prioritizing risk, LiveReview aims to reduce incidents, rollbacks, and the workload of senior reviewers, ultimately improving overall code quality and reducing the time spent on code reviews.

Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at dev.to →

More in Tech

Drafting Rights: Assign Every Doc Section to a Model or a Named Human

Most documentation pipelines fail on accountability rather than prose quality, because nobody can point at a sentence and name the human who approved it.

  • Models can draft only current behavior sections
  • Future functionality promises prohibited for models
  • Manifest file docs-rights.yml enforces tiers

Python: Call Multiple APIs in Parallel with asyncio.gather and Semaphore in 10 Minutes

When you need to call dozens of APIs — syncing data, crawling prices, firing webhooks — a sequential for loop adds up the waiting time of every request: 20 requests × 1 second = 20 seconds.

  • asyncio.gather enables concurrent execution of coroutines in a single event loop.
  • Semaphore limits concurrent requests to avoid overwhelming partner APIs.
  • Using returnexceptions=True allows handling of errors while returning a list of results.

Your Timer Is a Wall Clock, Not a Stopwatch: Four Elapsed-Time Myths and a Monotonic Probe

How long did that batch job take? Simple question. Your logs probably answer it wrong. I kept seeing two numbers for one run. The job log said eleven minutes. The dashboard said forty-two.

  • Wall clock measures what time it is, not how long a task took
  • Monotonic clock never goes backward, suitable for measuring durations
  • Timeouts indicate local socket issue, not remote task completion

More from Saturday 7 November →