agenterr vs glitchtip
Lighter than the lightweight Sentry alternative.
GlitchTip is the honest answer to "Sentry, but self-hostable" — same SDKs, a fraction of the stack. Agenterr starts from a different place: your services already emit logs, so error tracking can be one small process that reads them, instead of an app stack waiting for SDK events.
The stack difference
GlitchTip self-hosting runs a web app, Postgres, Redis, and background workers. Agenterr is one process and one data file, idling around 25MB — small enough to sit next to the apps it watches on the same $5 VPS.
The ingest difference
GlitchTip captures errors through Sentry SDKs in your code. Agenterr detects them in the structured logs you already write — no SDK to add, and warnings and context lines arrive with the errors. Sentry-SDK compatibility is on the agenterr roadmap.
The agent difference
GlitchTip exposes a REST API built for its own UI. Agenterr serves issues over MCP as the primary interface — your coding agent lists open issues, pulls one apart, and marks it resolved without you copying anything into the chat.
When GlitchTip is the right choice
You want drop-in Sentry SDK compatibility today, uptime monitoring included, or a proven Django stack your team already knows how to operate. Agenterr trades that maturity for weight and agent access.
 agenterr | GlitchTip | |
|---|---|---|
| Self-hosting | one process, one file | app + Postgres + Redis + workers |
| How errors arrive | structured logs (OTLP/JSON) | Sentry SDKs |
| Agent access | agent-first, MCP built in | REST API |
| Pricing | free self-host | per-event tiers |
| License | AGPL-3.0 | MIT |