Open source · Apache-2.0 · Your data stays yours

Chat with your company.
It remembers, and it does the work.

BrainOutput is a chat-first AI company: projects, tasks, documents and follow-ups — executed on free, local, or your own AI models. You approve anything before it goes out.

Evaluate for 7 days — no card Self-host it — Apache-2.0

7-day evaluation, no card · paid plans on your own EU VPS from €49/month · Free OpenCode models included · en / fr / de

Prefer a guided start? Book the AI Diagnostic →

380+ automated tests · EU-hosted · Apache-2.0, every line inspectable

your-workspace.trial.brainoutput.com
🏢 Your company
your models · your keys
💬 Chat
🗂 Work
⚙
Projects
q3-invoices3
website-refresh1
Chats
printer on vlan 3?
Chase every unpaid invoice this week.
BrainOutput
Mission · approved
3 unpaid invoices found (mail, read-only) — follow-ups to draft and send.
graph: single · worker · approvals: send-email
Run · live
✓ read mailbox … drafting 2/3
OpenCode Zen · free · sources shown per step
Projects list in BrainOutput: three client projects with task counts and status
Costs view in BrainOutput: inference ledger, budget and providers
What it takes off your plate

Real work, done — not another chatbot window

✉️

Your Alter handles mail

Reads only what you authorize, finds every thread waiting on you, and drafts replies in your tone. Nothing sends without your approval.

📄

Real documents

Quotes, brochures, briefs — delivered as actual files (PDF included), with your pictures fetched from your own website. Preview and download in one click.

🗂

Projects that remember

Every project keeps its conversations, tasks and results together. Ask for work in plain words — it drafts the mission, you approve, it runs in the background.

How it works

Up and running in an afternoon

1

Tell it about your company

A short conversation — name, what you do, your website. Connect mail, calendar or drives; everything starts read-only.

2

Ask for work in plain words

"Create the brochure", "chase last week's leads" — BrainOutput drafts a mission with the right agent and the right model.

3

Review and approve

Drafts wait for you. Consequential actions stop at a human gate — that's how the system is built, not a setting.

Why it's cheaper and safer

Your models. Your keys.
Your rules.

  • No per-token markup — free OpenCode models out of the box, your own provider key, or local models on your hardware.
  • Read-only by default — new connections can look, never touch. You grant each action explicitly.
  • Routines, not heartbeats — the daily digest and the regulation watch run on a schedule you see, never on hidden agent loops.
  • Full audit trail — every action records the model, the permission, the approval and its sources.
  • Open source (Apache-2.0) — inspect every line, or run it entirely on your own machine.
$ node bo-community.mjs serve
$ node bo-community.mjs ask "why is the printer offline?"

→ free model health-checked & connected (OpenCode Zen)
→ answer with sources — model: free, cost: €0
→ thread saved — open it anytime, it remembers
Deploy it your way

Hosted, Private Box, or Hybrid

Most AI tools lock you into one model: SaaS-only, or self-host-only. BrainOutput gives you all three — and you can switch or mix as your needs change.

☁️

Hosted

We run it for you on your own EU-hosted VPS — updates, backups and monitoring included. Live in about a minute, from €49/month after a free 7-day trial.

🖥️

Private Box

Run it entirely on your own hardware — an office server, a homelab, or on-premises. Open source under Apache-2.0; your data never leaves the building.

🔀

Hybrid

The best of both: a hosted workspace in the EU, with local models on your hardware for private work — or Enterprise in your own cloud project. Mix per role, per mission.

Pricing

Simple, honest pricing

Every hosted plan runs on your own VPS (virtual private server) — dedicated resources, exclusively your organization, from €49/month. The 7-day, no-card evaluation runs the full paid workspace on a small trial workspace; when you convert, your work moves with you. The subscription covers orchestration, updates, backups according to plan, and support. Model usage comes from free models, local models, your existing subscriptions, your own API keys, or an optional managed inference service.

Every paid plan runs on its own VPS (virtual private server). Its application runtime, data volume, credentials, queues, and active jobs are not shared with another customer. Plan numbers are allowances we enforce, not capacity claims.

Community €0 forever — the open-source runtime on your machine · self-host free

MOST POPULAR

Private Starter

€49 /month
  • Your own VPS (virtual private server)
  • 2 vCPU · 4 GB RAM · 40 GB storage
  • 20 TB traffic · Linux
  • Up to 5 people · 3 Work Twins
  • 2 simultaneous missions
  • Routines, approval gates & audit trail
  • Standard backups · email support
  • 7-day trial to your own VPS — no card
Request this plan Evaluate for 7 days — no card

Private Team

€99 /month
  • Your own VPS (virtual private server)
  • 4 vCPU · 8 GB RAM · 80 GB storage
  • 20 TB traffic · Linux
  • Up to 15 people · 10 Work Twins
  • 6 simultaneous missions
  • Longer audit retention
  • Enhanced backups · priority support
Request this plan Evaluate for 7 days — no card

Private Business

€199 /month
  • Your own VPS (virtual private server)
  • 8 vCPU · 16 GB RAM · 160 GB storage
  • 20 TB traffic · Linux
  • Up to 50 people · 30 Work Twins
  • 15 simultaneous missions
  • Advanced permissions
  • Extended backups · priority restore & migration support
Request this plan Evaluate for 7 days — no card

Enterprise On request — your own dedicated-CPU servers, your own cloud project, or on-premises · talk to us

The 7-day free trial runs on a small trial workspace (0.5 vCPU · 512 MB RAM · 5 GB storage) — no card. Every paid plan is your own VPS: dedicated resources, exclusively your organization, and your trial data moves with you when you convert. Plans differ by people, Work Twins, simultaneous missions, storage, backups and support.

Services

Other AI services

Fixed-scope offerings, separate from the plan ladder — they are not plans and are priced on request.

Diagnostic & Implementation — fixed scope and duration, on request · talk to us

BrainOutput ERP and further offerings will appear here when they ship.

Evaluate for 7 days — no card.

Every paid plan runs on your own EU VPS, from €49/month. Start with a 7-day evaluation on a small trial workspace — no card, and your work moves with you when you convert.

Evaluate for 7 days — no card Book the AI Diagnostic