Show HN: Sherlog Canvas – AI powered notebooks for debugging incidents
Why Sherlog? When an alert fires, you end up flipping between logs, dashboards, code, tickets, chat—losing context and precious time. Sherlog gives you a single canvas to:
Upload logs or connect to running docker containers (or kubernetes) (plain text, multiline, logcat, etc.) and analyze the logs and metrics
Run SQL queries against your database
Execute code snippets
Link GitHub Issues (or your ticket tracker)
Annotate hypotheses, build timelines, write notes
All cell types (logs, metrics, SQL, code, issues, CI/CD steps, etc.) are powered by MCPs, so you can interact manually with each integration—or let the Sherlog AI generate, execute, and refine cells automatically based on your queries.
Everything runs locally (via Docker), stores data locally, and makes external API calls only for the LLMs to openrouter. It’s open-sourced and available on github.
Current alpha features:
Interactive notebook UI
AI‑assisted summaries & root‑cause suggestions
Multi‑type cells backed by MCP for direct integration
Smart AI agents that correlate events across logs, metrics, and code
Roadmap:
MCP connectors: Datadog, Prometheus, Sentry, Jira, GitHub Actions
Mobile‑focused log support (Android/iOS crash analysis) (We are mobile engineers so this is personal itch we want to scratch)
Collaborative, real‑time canvases for team investigations
We built Sherlog because we noticed that come an incident or a bug we needed to gather information across multiple data sources/ tabs and often were using ChatGPT or Claude for generating queries for them. We just wanted to build an interface that would allow us to collect everything at one place and do triaging and investigation quickly and easily.
https://github.com/GetSherlog/Canvas https://getsherlog.com
Demo video - https://youtu.be/80c5J3zAZ5c
Would love to hear what’s missing, confusing, or downright broken!
No comments yet