AppSignal's MCP server hands Claude, Cursor, or Zed your real errors, traces, and the deploy that shipped them. AI writes the fix; you review the diff.
Free 30 days.
Host LLMs in Production With On-Demand GPUs
NVIDIA L4 GPUs. 5-second cold starts. Scale to zero when idle.
Deploy your model, get an endpoint, pay only for compute time. No GPU provisioning or infrastructure management required.
...It compiled without errors, but the emulator takes too long too long to load. I'm not sure if their testing tools even work properly, but the code is in a folder in the files section. I will be updating the project soon.