v8.5 — Now with Custom Agents & Workflow Builder

Your Private
AI Workspace

One platform replacing the 5+ disconnected AI tools your team already pays for.

Teams today use ChatGPT, Copilot, Notion AI, Perplexity, ElevenLabs and more — each with its own login, its own data policy, and zero visibility into actual usage. Hubrix replaces them all.

Continue with Google Continue with Microsoft

Or sign in with email & password · Free trial · No credit card required

26
API Routers
981×
Cache Speedup
3
AI Providers
111/111
Tests Passing
EU
Hosted + On-Prem
The Problem

Your Team Already Uses AI.
But It's Fragmented.

The average knowledge-worker team pays for 5–7 separate AI subscriptions. No shared context. No cost visibility. No data governance.

What your team pays today

💬 ChatGPT Team€20/user
🤖 Copilot M365€30/user
📝 Notion AI€16/user
🔍 Perplexity Pro€20/user
🎤 ElevenLabs€22/team
🌐 DeepL Pro€25/team
No data governance
No cost visibility
No shared context
€130+ per user / month · fragmented · ungoverned

Replaced by Hubrix

💬Smart Chat (Claude + GPT + Gemini)included
📄Document Intelligence (RAG)included
🔬Deep Research (20+ sources)included
🤖Custom AI Agentsincluded
⚡Workflow Automationincluded
🎤Voice Chat (STT + TTS)included
€34.90 per user / month · governed · visible
Platform Features

12 AI Capabilities.
One Platform.

One authentication layer. One credit system. One admin panel. Zero fragmentation.

💬
Smart Chat + Web Search
Multi-model AI chat with 9 specialist agents (Legal, Marketing, Developer, Analyst, Researcher, Translator, and more). Real-time web search via Serper. SSE streaming.
Claude Sonnet 4.6GPT-5.4 MiniGemini 2.5
📄
Document Intelligence (RAG)
Upload PDFs, Word, Excel. Import from Google Drive, OneDrive, SharePoint. Ask questions and get precise answers using pgvector semantic search with source citations.
pgvectorPDF · DOCX · XLSXSource citations
🔬
Deep Research
AI autonomously searches 20+ web sources, synthesizes findings, and produces a cited structured report. Export to Word. Three depth levels (5–20+ sources).
20+ sourcesCited reportsDOCX export
🤖
Custom Agent Creator NEW
Build your own AI specialists: custom name, emoji, system prompt, model, temperature, and tools. Share with your team. Appears alongside built-in agents everywhere.
Custom promptsTeam sharingPlan limits
⚡
Workflow Builder NEW
Chain multiple agents into sequential pipelines. Each step gets output from the previous. Run with variables. Live SSE progress. Automate your team's most repetitive multi-step AI tasks.
Multi-stepVariablesLive progress
📊
Bulk Processing
Apply any AI prompt to thousands of Excel or CSV rows automatically. Background processing via ARQ queue. Download results as formatted Excel. Handles 10,000+ rows.
Excel · CSVAsync queueThousands of rows
🗂️
Team Workspace
Shared document library with folder structure, workspace memory injection, member management, and full workspace activity logging. Collaborate on AI-powered knowledge bases.
Shared memoryAudit logFolders
🎤
Voice Chat
Real-time voice input with Whisper STT and natural voice output via ElevenLabs TTS (21 voices). No separate app — works directly in your browser. Works with all agents.
Whisper STT21 voicesBrowser-native
🔑
REST API Access
Every Hubrix feature available via REST API with Bearer token auth. Integrate into your CRM, ERP or internal tools. 150+ endpoints documented at /api-docs.
Bearer auth150+ endpointsAll features
Workflow Builder

Automate Your Team's
Multi-Step AI Tasks

Chain agents into pipelines. Each step passes its output to the next. Set up once, run with variables. Live progress as each step completes.

Example: Market Research Pipeline

1
🔬 Research Agent
Research the competitive landscape for {topic} in the {market} market. Find the top 5 players, their pricing, and key differentiators.
↓
2
💰 Financial Analyst
Analyze the competitive data below and identify market gaps and pricing opportunities: {previous_output}
↓
3
✍️ Copywriter
Write an executive summary and strategic recommendation based on this analysis: {previous_output}
✅ Workflow complete · 3 steps · 45 credits used · Report ready

What you can automate

📋 Research → Analysis → Report
3 steps. Runs in minutes. Replaces 3 hours of manual work.
📝 Brief → Draft → Edit → Localize
4-step content pipeline with your custom brand agents.
⚖️ Contract → Risk → Summary → Email
Legal review pipeline. Custom legal agent. One click.
📊 Data → Insights → Slide deck
Feed in Excel data, get a structured insights report out.
Plan limits: Starter = 5 workflows × 3 steps · Professional = 20 × 10 · Enterprise = unlimited
Single Sign-On

Sign In With
Google or Microsoft

No extra passwords. Connect your existing Google or Microsoft account in one click. Enterprise users get seamless access with their organization accounts.

  • ✓ Auto-link existing Hubrix accounts by email
  • ✓ Connect/disconnect providers in Settings
  • ✓ Separate auth for Data Connectors (Drive scope)
  • ✓ Works alongside email/password login
Login with Google Login with Microsoft
Data Connectors

Import Files from
Cloud Storage

Browse and import files from Google Drive, OneDrive, and SharePoint directly into your AI document library for instant semantic search.

Google Drive
Browse folders, import PDFs, Docs, Sheets and more. Google native formats auto-exported.
Browse→ Select→ Import to RAG
OneDrive
Connect your Microsoft account and access personal or work files stored in OneDrive.
Browse→ Select→ Import to RAG
🏢
SharePoint
Access your organization's SharePoint sites, document libraries and team files.
Sites→ Libraries→ Import to RAG

Supported: PDF · DOCX · XLSX · PPTX · TXT · CSV · JSON · Google Docs/Sheets/Slides

Semantic Cache

Instant Responses
For Repeated Work

Your team asks similar questions every day. Our three-level semantic cache stores and retrieves answers fast — reducing AI costs and response times dramatically.

Level 1 — Redis exact match (SHA256) — cross-model enabled

Level 2 — pgvector cosine similarity (threshold: 0.82)

Level 3 — Smart TTL for time-sensitive queries

981×
Faster on L1 hit
~11ms
L1 exact hit (measured)
100%
Margin on cache hit
Response Time — Production Measured
Cold AI call10,787ms
Level 2 Semantic533ms
Level 1 Exact ⚡11ms
Admin → ⚡ Cache Tab
247
Cache Hits
€0.82
Cost Saved
Competitive Landscape

How Hubrix Compares

The only platform combining multi-model AI, on-premise deployment, workflow automation, and enterprise cost controls — at SME pricing.

Feature Microsoft Copilot ChatGPT Team Notion AI Hubrix
Multi-model (Claude+GPT+Gemini)GPT-4o onlyGPT onlyGPT only✓ All 3 providers
Custom AI AgentsNoNoNo✓ Build your own
Workflow automation (multi-step)NoNoNo✓ Pipeline builder
On-premise / self-hostedNoNoNo✓ Docker Compose
Deep research (20+ sources)NoLimitedNo✓ Cited reports
Bulk processing (Excel/CSV)NoNoNo✓ Thousands of rows
Company budget controlsPartialNoNo✓ Hard EUR cap
Per-user cost visibilityPartialNoNo✓ Per-model detail
EU-hosted / GDPR-consciousPartialUS onlyUS only✓ Hetzner EU + on-prem
Price (team of 10)€360+/mo€200/mo€160/mo€349/mo

* Pricing and features shown are illustrative comparisons based on publicly available information. Competitor offerings may change. Contact us for a detailed comparison for your team size.

Headless API

Integrate AI Into
Your Own Apps

Use Hubrix as an AI backend for your CRM, ERP or internal tools. Full REST API with Bearer token authentication. 150+ endpoints documented.

🔑Generate API keys in Settings — max 10 per user
📖Full docs at /api-docs — Python, JS, curl examples
🔒Scoped access: chat · rag · translate · agent · bulk
📊Per-key credit analytics and usage tracking
View API Documentation →
# Python — Chat via API Key import requests API_KEY = "sk_live_ocean_YOUR_KEY" BASE = "https://app.hubrix.ai/api/v1" response = requests.post( f"{BASE}/chat", headers={"Authorization": f"Bearer {API_KEY}"}, json={ "message": "Summarize this contract", "agent": "legal", "model": "claude-sonnet-4-6" } ) print(response.json()["response"]) # ← Instant if cached, AI response if not
Pricing

Simple, Transparent Pricing

All plans include access to all features. Credits consumed per AI action — different features use different amounts.

Starter
For small teams getting started
€149 /mo
Up to 3 users · 2,000 credits/month
  • Smart Chat + Web Search
  • Document Intelligence (RAG)
  • Translate Studio (21 voices)
  • SSO (Google + Microsoft)
  • Data Connectors
  • 5 Custom Agents · 5 Workflows
  • Admin Dashboard
  • Email support
Get Started

€1.66/user/day

⭐ Most Popular
Professional
For growing teams
€349 /mo
Up to 10 users · 6,000 credits/month
  • Everything in Starter
  • Deep Research (20+ sources)
  • Bulk Processing (Excel/CSV)
  • AI Agent Runner
  • 20 Custom Agents · 20 Workflows
  • REST API + API Keys
  • Semantic Cache Analytics
  • Priority support
Get Started

€1.16/user/day

Enterprise
Full control & customization
Custom
Unlimited users · On-premise option
  • Everything in Professional
  • On-premise deployment (Docker)
  • Unlimited Custom Agents + Workflows
  • Custom markup + EUR budget mode
  • GDPR DPA agreement
  • SLA guarantee
  • Dedicated onboarding
Contact Sales
💡
Break-even: 3 minutes of time saved per person per working day. At €349/month for 10 people (€1.16/user/day) — if each team member saves just 1 hour per month using Hubrix instead of manual research or writing, the subscription has already paid for itself.
Start Free Trial →

💡 AI provider costs (Anthropic, OpenAI, Google) are included in your credit balance — no separate bills.

Why Hubrix

Built for Teams That
Need Real Governance

Not a chatbot wrapper. A production-grade AI workspace with enterprise controls from day one.

01
Stronger Data Control
Self-hosted deployment option via Docker Compose. Your data, documents, and session history stay on your own infrastructure. AI processing calls go to external providers — only the platform runs on your servers.
02
Full Cost Transparency
Real-time tracking of every AI token used. Per-user daily limits. Monthly EUR budget cap per company. Semantic cache shows exactly how much you save. No surprise bills.
03
Model Flexibility
Not locked into one AI provider. Switch between Claude Sonnet 4.6, GPT-5.4 Mini, and Gemini 2.5 Flash per task. Same credit cost regardless of model. Always the latest versions.
04
SSO & Identity
Sign in with Google or Microsoft. 4-tier role model: System Admin → Company Admin → Member → Individual. Token blacklist on logout. Row-Level Security on 13 DB tables.
05
Local Support — Netherlands
Based in Poortugaal, South Holland. Hubrix Consulting VOF · KVK 84553081. Hosted on Hetzner EU infrastructure. GDPR-conscious deployment options. SLA available.
06
Fully Tested & Operational
111/111 API tests + 107 browser tests at 100% pass rate. 26 API routers. 46 database tables. Grafana monitoring. Daily automated backups. Not a prototype — fully operational.
⚡ FastAPI + Python 3.12
🐘 PostgreSQL 16 + pgvector
⚡ Redis Cache + ARQ Queue
🔐 JWT + OAuth 2.0
📊 Grafana Monitoring
🐳 Docker On-premise
☁️ Google Drive API
☁️ Microsoft Graph API

Ready to Replace Your
Fragmented AI Stack?

Free trial · No credit card required · Available for pilot deployments today.

🚀 Start Free Trial 📅 Book a Demo