Bring your own AI: run Junyr's AI on your company's own server

Your company administrator connects the Junyr Suite's realtime AI to an LLM server your company controls, once, for everyone. Your data never leaves your machine, on every plan.

Bring your own AI: run Junyr's AI on your company's own server

In short: your company administrator points every realtime AI feature at an LLM server your company controls, once, in the company AI settings. When it's reachable, summaries, search, smart replies and Junyr Agent chat run on your hardware instead of the cloud, for everyone in the company, and your data never leaves your machine. It's included, on every plan.

---

The independence problem with "AI for business"

Most AI products ask you to make a trade: useful automation, or control over your data, pick one. For an inbox full of salaries, IBANs, contracts and client secrets, that's not a real choice. So we built the other option into the Junyr Suite, the independent AI operating system for your business.

In the company AI settings (Settings → AI and models), a company administrator gives Junyr the address of a model server your company controls. This is set up once, at the company level, and from that moment every eligible AI request runs there for the whole team.

---

Any server, your company's choice

Junyr speaks the OpenAI-compatible API, so it works with whatever your company already runs:

• Ollama, LM Studio, vLLM, or any OpenAI-compatible gateway your company chooses
• On your LAN, your Tailscale network, or your own datacenter

A built-in connection test confirms reachability and discovers the models the server exposes. Per-modality capability probes run a real inference for each capability you care about: chat (text→text), vision (image→text), transcription (speech→text) and embeddings, and report whether each one works, with a sample and a latency reading.

---

What runs locally

When the company's server is online, Junyr routes essentially all of its realtime AI to it, for every user in the company:

• Compose review and reformatting before you send
• AI mailbox search ("find the renewal quote from spring")
• Smart actions on an open email
• Junyr Agent chat, Ask Junyr and project discovery
• Workflows, and even the nightly Reflexions pipeline

There's no separate "local mode" to babysit. It's the same product for everyone; the inference just happens on your box.

---

True independence, not marketing independence

This is the part that matters. Junyr has three per-inbox confidentiality tiers: Totale (no cloud AI at all), Sécurisée (AI with personal data masked first), and Simple. A reachable company server changes what's possible for every mailbox:

• It lifts the Totale block. Your most confidential mailboxes get summaries, smart replies and search, because the request physically never leaves your hardware.
• It skips the Sécurisée masking step for calls that run locally. Salaries, IBANs and contracts stay intact, because there's no third party to hide them from.

And when the company's private server is offline, Junyr fails safe: confidential-tier calls return an error rather than quietly falling back to the cloud. Nothing leaks because the system would rather do nothing than betray the tier your company chose.

---

No hardware? Your administrator can bring a company cloud key instead

Running a model server isn't for every company. If your administrator would rather use a commercial model but still own the relationship, they can plug in the company's own Gemini, Claude or Mistral API key (encrypted at rest), for the whole team. Your company controls the provider and the bill.

One honest caveat: a company cloud key is still cloud. It's gated exactly like Junyr's platform cloud: Totale blocks it, Sécurisée masks personal data first. Only the local server is fully independent. We don't blur that line.

---

Private by design

• SSRF-guarded endpoint validation blocks cloud-metadata addresses while still allowing LAN and Tailscale hosts
• The connection credential is write-only and stored encrypted, never sent back to the browser
• The setting is company-wide: once your administrator configures it, it applies to everyone, with nothing for individual users to set up

---

Included on every plan

This isn't an "Enterprise" upsell. Bring-your-own AI is part of the all-included Junyr Société plan and the additional user seats stacked on top of it, with no separate add-on to buy. Hosted in Europe by default, on your company's own hardware when you want it: that's the whole point.

Your company administrator sets it up once in Settings → AI and models and the whole team runs on AI your company actually controls. Want the broader picture? See how the Junyr Suite keeps your data independent or read about the three confidentiality tiers that govern every AI call.

---

FAQ

Can we run Junyr's AI entirely on our own hardware?

Yes. Once a company administrator points Junyr at any OpenAI-compatible server your company controls (Ollama, LM Studio, vLLM or a gateway on your LAN, Tailscale network or datacenter), every eligible realtime AI request from every user runs there. When the server is reachable, your data physically never leaves your machine.

Does our company's bring-your-own AI cost extra?

No. It's included in the all-included Junyr Société plan (179 €/month HT, first user included) and every additional 39 €/user/month seat, with no add-on catalogue. See the pricing page for the full breakdown.

What happens to our most confidential mailboxes?

A reachable company server lifts the Totale tier block and skips the Sécurisée masking step for calls that run locally, because the request never reaches a third party. If the server goes offline, Junyr fails safe: confidential-tier calls return an error rather than quietly falling back to the cloud.

Who can set up the company's LLM server?

Only a company administrator (an admin or super-admin role) can reach the company AI settings and connect a local server or a cloud key. Once it's configured, it applies to every user automatically, with nothing for individual team members to set up. Without a configured server, your company can still plug in its own Gemini, Claude or Mistral API key, encrypted at rest, and keep control of the provider and the bill. One caveat: a cloud key is still cloud, Totale blocks it and Sécurisée masks personal data first. Only a local server is fully independent.