First run

From install,
to a working cluster.

The install command returned. The daemons are up. This page covers the next twenty minutes — claiming the admin account, activating your license, loading your first knowledge base, and the first useful chat against your own data.

1. Verify the cluster is running.

On the host you installed onto, confirm the service status:

sudo systemctl status eldric-aios

You want to see active (running). If anything's red, jump to the troubleshooting page. Open https://<your-host>/chat in a browser. You should see the Eldric chat shell with a first-run signup card.

2. Become the first admin.

Eldric ships with no default password. The very first user to sign up on a fresh cluster becomes the system administrator — full rights over tenants, users, the license, and configuration. Make sure the right person on your team signs up first; subsequent signups land as regular users in the default tenant.

The first-run card prompts for an email and a password. After confirming, you land in the chat as admin with the gear icon revealing the Admin Console.

If the cluster is publicly reachable on the internet, disable open signup from the Admin Console immediately after claiming the admin account, and add team members manually from the Users page. Open signup on a public host is a fastest-finger race for the admin role.

3. Activate your license (optional).

The free tier ships with the full kernel: one controller, one router, two workers, all features. Skip this step if that's enough. To scale beyond, request a license file from license@core.at and activate it via the Admin Console:

Admin Console → License → Upload license file.

Or via the API:

curl -X POST https://<your-host>/api/v1/license/activate \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"license_file": "/etc/eldric/license.json"}'

The controller verifies the Ed25519 signature and applies the new limits on the spot. No restart.

4. Add your first knowledge base.

The whole point of running Eldric on your own hardware is that you can let it learn from your data. From the Admin Console:

Admin Console → Knowledge Bases → New KB.

Name it something obvious (“company-handbook”, “clinical-guidelines”, “contracts-2026”). Upload documents — PDF, DOCX, Markdown, plain text, HTML. The data worker chunks them, embeds them, and stores the vectors locally. Done in a few minutes for hundreds of pages; longer for thousands.

Back in the chat, switch the active KB from the chat header and ask questions of your own documents. Sources are cited inline.

5. First chat.

Try the simplest case first — just type a question. The chat streams the answer, citing your KB documents in the margin. Then push further:

• Try a web search. Type /websearch <query> to invoke the searcher agent. Results land back in the chat with source links.
• Attach a file. Click the 📎 icon, drop a PDF or a chart. Eldric ingests it inline for that conversation.
• Hand off to an agent. Type /agent coder review this to delegate. The coder agent reads, analyses and reports back.

Next steps.

The platform has more depth than five steps cover. Three good directions from here:

• Developers — the OpenAI-compatible API, plugin authoring, webhook subscribers.
• Features catalogue — eighteen domains, every shipped feature.
• Use cases — what people in your sector typically run on Eldric.

Stuck on a specific step? Common failure modes are collected in the install troubleshooting page. Or write to office@eldric.ai.