SentinelReady

Run SentinelReady Community — free

The same triage engine paying customers run. Community is a capacity limit, not a cut-down build — the AI, the pattern learning, the fail-open behaviour and the delivery logic are identical. No card, no licence key, no expiry.

1Ask for access

Email support@sentinelready.io and say you'd like to run Community. You'll get repository and container registry access back, usually within one business day.

We set people up by hand rather than shipping a bare download. It means a real person knows you're running it and you're not debugging your first install alone — the same way paying customers are onboarded.

2Pull and run

Once your access lands:

docker login ghcr.io -u YOUR_GITHUB_USERNAME
# paste the access token from your setup email when prompted

git clone https://github.com/browningjc97/sentinelready-docker
cd sentinelready-docker
cp sentinelready.yaml.example sentinelready.yaml
docker compose up -d

No licence key to enter — Community is what SentinelReady runs as by default. First startup takes 3–5 minutes while it pulls Ollama and the llama3.1 model; after that, seconds.

Budget about 8 GB of RAM, most of it the model. Already running Ollama? Use docker-compose.external-ollama.yml and point at yours instead of running a second copy.

3Point one alert source at it

Prometheus/Alertmanager, as an example — add a receiver:

receivers:
  - name: sentinelready
    webhook_configs:
      - url: 'http://your-sentinelready-host:8000/webhook/auto'
        send_resolved: true

Send one real alert and watch what comes back. There are dedicated endpoints per tool (/webhook/grafana, /webhook/datadog and so on) — or point anything at /webhook/auto and SentinelReady works out the field mapping itself the first time it sees that payload shape.

What you get, and where it stops

The full engine. No AI provider is licence-gated — run local Ollama, or bring your own Claude or OpenAI key, on the free tier. Pattern learning, causal correlation, fail-open delivery and severity handling all work exactly as they do on Pro.

The limits: 50 learned patterns, one user, and five sitrep digests. After those five, SentinelReady keeps sending a weekly heartbeat — how many alerts it handled, how many it escalated, how many it kept off your phone. That continues forever, at no cost, so a working install never goes silent on you.

Support is GitHub issues. Community is genuinely free and doesn't expire — but it isn't a support contract. If SentinelReady becomes something you depend on, Pro lifts the pattern cap, adds unlimited delivery targets and multiple users, and comes with real support.