Changelog #16: Fix errors with AI SRE

You shouldn't have to fix every error yourself. Hand any error to AI SRE and get back a pull request with the fix, from your desk or your phone.

Here's how to fix errors with AI SRE.

Turn an error into a pull request

See a new error? Hand it to AI SRE. It analyses the stack trace, related logs, environment, and recent releases to find the root cause. Connect your GitHub or GitLab and get a pull request with the fix. Review, merge, done.

Hand your next error to AI SRE.

Prefer to drive the fix locally?

Use our MCP server or copy a ready-to-use prompt to investigate any error. Better Stack MCP reads errors, stack traces, logs, releases, and all other data AI SRE can. Our pre-made prompts contain all the important details out-of-the-box and they are customizable. Once done with analysis, just ask your local agent to fix the error.

Install Better Stack MCP server or use our AI prompts.

Fix errors on the go

Got a notification for a critical error? Start a new AI SRE chat from the Better Stack iOS and Android apps. Ask it to list recent errors, pick one, then tell it to investigate and ship a fix as a pull request.

Need to check back on AI SRE? Jump back into the same chat in the app or on Better Stack.

Download our iOS and Android apps.

Not using Better Stack Error tracking yet?

Start sending errors to Better Stack in 5 minutes. Simply swap a token in your Sentry SDK and you're done. No re-integration needed. Each error costs a sixth of Sentry's price.

Start using Sentry-compatible error tracking.

---

Fresh community content

Observe Claude Code with OpenTelemetry

Claude Code exports its own usage data over OpenTelemetry: sessions, tokens, cost, prompts, tool calls, and errors. Point the export at Better Stack and see what your team’s AI actually does, what it costs, and where it fails.

What’s New in OpenCV 5

OpenCV’s first major release since 2018 ships a new DNN engine, real-time YOLO object detection, diffusion-based inpainting, and image captioning with PaliGemma. Computer vision’s standard toolkit just caught up with the AI era.

This Tiny Engine Runs Impossibly Big AI Models Locally! (colibrì)

Can a 744-billion-parameter model run on consumer hardware? Colibrì is a tiny pure-C engine that streams GLM-5.2 straight off your SSD, tested on a MacBook and an RTX 5090. What’s holding local AI back isn’t your GPU.

Apple Just Built WSL for the Mac (Container Machines)

Lightweight, persistent Linux environments on your Mac, built on Apple’s Container project and optimized for Apple Silicon. WSL-style dev workflows finally arrive on macOS.