Connect PrivatAI to Odysseus

In May 2026, Felix Kjellberg — better known to 110+ million subscribers as PewDiePie — released Odysseus, a free, open-source, self-hosted AI workspace. It bundles chat, agents, deep research, a document editor and email into one interface that runs on your own hardware. Within weeks it had crossed 77,000 stars on GitHub.

The philosophy behind Odysseus lines up almost exactly with ours: your data should stay yours. No accounts, no telemetry, no sending every keystroke to a big-tech server.

But there's a gap. Odysseus shines when you run models locally through Ollama, llama.cpp, LM Studio or vLLM — yet not everyone has a GPU that can host a capable model. That's where Odysseus lets you plug in a cloud API instead. The catch: most cloud APIs are US-hosted, which pulls your prompts right back out of the EU and into exactly the kind of data-collection pipeline Odysseus was built to avoid.

PrivatAI closes that gap. We're OpenAI-compatible, so Odysseus can use us as a drop-in provider — and every request stays on EU infrastructure.

flowchart LR
    A[Odysseus Workspace] -->|local| B[Your GPU]
    A -->|cloud, default| C[US Provider]
    A -->|cloud, with PrivatAI| D[PrivatAI EU]
    D --> E[Private EU Model]

Why pair them?

  • No local GPU needed — run large models without buying hardware.
  • Stays in the EU — no detour through US servers, no US CLOUD Act exposure.
  • No logs, no training on your data — same promise as running locally, just hosted.
  • Zero code changes — Odysseus already speaks the OpenAI protocol.

Prerequisites

Step 1: Create an API key

Sign in to PrivatAI, open the dashboard, and create an API key. Copy it immediately — it's shown only once.

Step 2: Add PrivatAI as a provider in Odysseus

In Odysseus, open the model / provider settings and add a new OpenAI-compatible provider. Point it at PrivatAI:

  • Base URL: https://privatai.com/api/v1
  • API key: privat_your_key
  • Model: gpt-oss-120b (available on any plan) or glm-5.2 (Professional)

If you prefer environment variables when launching the Odysseus container, the same values apply:

export OPENAI_API_BASE="https://privatai.com/api/v1"
export OPENAI_API_KEY="privat_your_key"

Step 3: Test it

Before wiring it into your agents, confirm the endpoint answers:

curl -X POST https://privatai.com/api/v1/chat/completions \
  -H "Authorization: Bearer privat_your_key" \
  -H "Content-Type: application/json" \
  -d '{"model": "gpt-oss-120b", "messages": [{"role": "user", "content": "Hello from Odysseus!"}]}'

If you get a reply, PrivatAI is live. You can now select it as the model in Odysseus chat, agents and deep research.

That's it

Odysseus gives you a polished, private workspace. PrivatAI gives it a capable model without a GPU and without leaving the EU. Two projects, one principle: your conversations belong to you — not us, not third parties.

Questions or feedback? Write to us at info@privatai.com.