opensre binary yourself — on a laptop, a shared VM, or a machine in your own network.
Which setup fits you?
Platform steps:
1. Install the CLI
sudo. The installer puts opensre somewhere already on your PATH when it can. If it can’t, it uses ~/.local/bin and prints how to add that to PATH.
Auto onboard: In a normal interactive terminal, install often starts opensre onboard when it finishes. In CI or scripts it skips that. To turn auto-onboard off:
gh if it’s missing (OpenSRE’s GitHub tools need it). Skip that with:
gh automatically.
Shorter overview: Install OpenSRE.
2. Open the shell
opensre … from a normal terminal without entering the shell.
3. Run onboarding
Inside the shell:- Your LLM (so OpenSRE can reason about incidents)
- Tool credentials (Datadog, Grafana, Slack, and so on)
- Optional messaging and reporting
Where credentials are saved
OpenSRE checks environment variables first. When onboarding saves a credential, it writes to an owner-only file at~/.opensre/credentials.json (mode 0600).
That keeps first-run setup from writing credentials to disk.
You can also provide a credential through your shell or project .env;
environment values take precedence. Never commit real credentials to .env or
source control.
To keep credentials only in the environment (nothing written to disk), set
OPENSRE_DISABLE_KEYRING=1.
4. Check your tools
Inside the shell:opensre (for example opensre integrations verify).
5. Run a first investigation
Inside the shell:Where to connect tools
The full tool list lives under Integrations, not Install. Start at Integrations overview, then open each provider page for keys, permissions, and examples.Related
- Quickstart — install + onboard + sample run with GIFs
- Investigations overview — how investigations work end to end
- Deployment — gateway and hosted options beyond a single laptop