nitpickr | AI code review, on your infrastructure

What it does

Simple to explain.

nitpickr keeps the review loop in GitHub and keeps the control with your team.

Inline comments, a summary, and a nitpickr / review check.

Start local, use Railway for uptime, or keep it on your own VPS.

Configuration lives in the repository, not inside a hosted dashboard.

Why self-host

Local Docker is the fastest way to try nitpickr. Railway is the simplest always-on path. A private VPS gives you the strongest runtime boundary.

Local Docker

The fastest evaluation loop for a real repository.

Private VPS

Keep logs, secrets, and model traffic inside infrastructure you control.

Railway

The simplest always-on setup when you want a stable webhook endpoint.

Control

nitpickr is source-available, configured at runtime, and shaped by the repository itself.

OPENAI_MODEL=<your-model>
OPENAI_BASE_URL=https://api.openai.com/v1
NITPICKR_PROMPT_OPTIMIZATION_MODE=balanced
.nitpickr.yml
AGENTS.md

Choose the model, keep the rules in Git, and inspect the code yourself.

FAQ

Is nitpickr open source?

nitpickr should be described as source-available. The code is published under Elastic License 2.0.

Why would I host this on my own VPS?

Because some teams want the runtime, logs, secrets, and model traffic inside their own operational boundary.

What can I control today?

The deployment location, the repository rules, and the model endpoint through OPENAI_MODEL and OPENAI_BASE_URL.