Automate 4K sky camera and meteor classification with Claude Code
Use the open-source openskycam project with an MCP server to let Claude Code run overnight workflows, review and classify detections, retrain models to filter false positives, and deliver curated results every morning after cleanup.
Reddit r/ClaudeAI/u/Terrible_Award_2124
Claude Codeopenskycamedmund-harnessCloudflareRaspberry PiNode.js
Added 2026-10-31 · originally published 2026-10-26
Steps
- 1Prepare an RTSP camera, a Raspberry Pi, a macOS or Linux machine with large disk storage, and a Cloudflare account with a domain.
- 2Run the installation command npm run setup -- --yes and configure the .env file as prompted, while manually creating the Cloudflare tunnel, logging into the app, and setting up API keys.
- 3Use the openskycam MCP server to let Claude Code automatically build and run the overnight workflow, processing video frames recorded by the 4K sky camera.
- 4Have Claude Code review, accept, or reject detection results, and automatically retrain the model based on the approved dataset when the false positive rate is high.
- 5Set up a cron job to have the system automatically filter out spam every morning and send curated detection results via iMessage.
Copy-ready prompt
Please reference the MCP server guide and skills from the openskycam project to help me connect my local 4K sky camera data source. Assist me in running the overnight workflow, reviewing detection results, and classifying meteors, planes, and fixed lights. If you find too many false positives, help me retrain the model based on my approved dataset, and automatically clean up junk detections each morning before sending me the curated results.
#AI Agent#Claude Code#Automated Workflow#Open Source
Related use cases
Summarized by AI from the original source - refer to the original post for details. Copyright belongs to the original author.