QWorker is a personal assistant
QWorker is a contextual AI engine deployed inside your own infrastructure. Employees get a modern assistant — you keep control of data, models and access.
Public demo. Don't paste personal or confidential data. Messages are not stored.
Zero prompting · Zero data leakage · Full context
Built for regulated European teams
- EU-hosted
- GDPR & AI Act
- SOC 2-ready
- On-prem · Air-gapped
Shadow AI is already inside your organization
Your employees are already using ChatGPT at work. Most pay for it privately. Some have pasted confidential data out of the building.
Three questions your CISO can't answer today
- Which AI models are your employees using?
- Where does the data they paste in actually end up?
- What have they already shared?
QWorker gives them the same experience — on your terms.
Think. See. Act.
Three capabilities working together to turn a task description into finished work — inside your own environment.
The brain
QWorker plans the task, breaks it into steps and checks its own work before delivering. Built on models you choose and approve.
The eyes
Reads the active document, screen or dataset. Understands tables, fields and structure — no manual context pasting.
The hands
Executes actions in your tools: fills forms, edits files, calls internal APIs. Every step is logged and reversible.
AI that sees what the user sees
QWorker reads the active work context — a spreadsheet, a CRM screen or a tender document. The user doesn't have to explain the whole task in a long prompt.
QWorker understands the context
It detects the active window, document type, fields and current task.
The user picks an action
For example: analyze, summarize, format, validate or generate.
The work runs locally
Output is delivered straight back into the user's environment.
See QWorker in action
Watch a real task run from prompt to finished output — entirely inside your perimeter.
Summarise 240 case documents and draft a memo with citations
- 01Plan
Breaks the task into 4 subtasks and lists the documents to index.
- 02Read context
Embeds and retrieves the 18 most relevant passages from the case folder.
- 03Generate draft
Assembles a memo with inline citations back to the source documents.
- 04Hand back
Delivers the draft into the case worker's editor — every step in the audit log.
Give QWorker a list. Get the work back.
Like a command center for agentic work. Multiple agents work in parallel on your tasks — inside your own systems, with your data, under your rules.
- Update employee handbook with new vacation lawDoneHR agenthr-docs·qworker/ferielov-20264 min ago+142 −38
- Pull Q3 numbers from ERP and build board reportWorkingFinance agentfinance/reporting·qworker/q3-boardrunning now
- Answer 12 new support tickets in ZendeskWorkingSupport agentsupport/inbox·qworker/tickets-batch-318running now
- Screen 38 applicants for warehouse workerQueuedRecruiting agenthr/recruiting·qworker/shortlist-lagerqueued
- Check supplier contracts for GDPR gapsQueuedLegal agentlegal/contracts·qworker/gdpr-reviewqueued
Multiple agents solve tasks at once — across HR, finance, support and legal.
Every task comes back with sources, diffs and an audit trail you can approve.
Runs on your own infrastructure. No data leaves the building.
A modern assistant for users. A control panel for IT and compliance.
Everything employees expect from a modern AI tool — plus the controls your security, legal and compliance teams need.
Any model, locked down
Switch between GPT-4, Claude, Gemini and open-source models — or restrict access to only the approved ones.
SSO and role-based access
Login via your IdP. Roles, groups and an audit log on every conversation — so you can answer 'who asked what, and when'.
Workspaces that keep data apart
Separate HR from IT. Search across your own history — never anyone else's.
Markdown, code and math
Markdown, LaTeX and syntax-highlighted code blocks. What employees already expect.
Pin great prompts
Lock in a 'legal assistant' or 'press release template' across the company. Same quality in every department.
Plugins and agents
Extend with retrieval, internal tools and agents tailored to your processes.
Built for workflows where data must not leak
Sales and CRM
Sales copy, account summaries and follow-ups directly inside the system.
Procurement and compliance
Analyze supplier material and contract terms without data leakage.
Defense and government
Run AI inside closed or air-gapped environments.
Enterprise teams
Train the AI on your own processes and automate repetitive work.
Consulting firms
A fast, easy-to-deploy AI assistant for analysis and quality assurance.
Same work — without sending your data across the Atlantic
Littlebird remembers what you see on screen. Coworker routes between models and runs agents across 50+ systems. QWorker does both — on your own infrastructure, in the EU, with local support and documented compliance.
Remembers what you've seen
Find decisions, documents and conversations across your apps — without having to re-explain the context.
Acts across systems
One request triggers actions in calendar, mail, CRM, case handling and documents — in a single flow.
Creates without starting over
Drafts for mail, notes and reports are anchored in your actual work, not in an empty prompt.
Meeting notes and follow-up
Transcribes, summarizes and extracts decisions and action items — automatically after every meeting.
Cloud chatbot
ChatGPT, Copilot
- Where does the AI run?Vendor cloud (US)
- Data residencyUnknown or shifting
- GDPR & AI ActGray area
- Remembers your contextNo
- Multi-model routingNo — single vendor
- Acts in your systemsNo
- Agents per departmentNo
- Air-gapped optionNo
- Trains on your dataOften yes
- Local supportNo
- Pricing modelPer user
- Suited for critical infraNo
Littlebird
littlebird.ai
- Where does the AI run?AWS, US-hosted
- Data residencyUSA
- GDPR & AI ActSelf-declared
- Remembers your contextYes — screen + meetings
- Multi-model routingLimited
- Acts in your systemsFew integrations
- Agents per departmentPersonal routines
- Air-gapped optionNo
- Trains on your dataNo, but data in cloud
- Local supportNo
- Pricing model$17–100 / mo / user
- Suited for critical infraNo
Coworker
coworker.ai
- Where does the AI run?US-hosted
- Data residencyUSA
- GDPR & AI ActSOC 2
- Remembers your contextYes — knowledge graph
- Multi-model routingOpen + closed models
- Acts in your systems50+ connectors (US)
- Agents per departmentAgent library
- Air-gapped optionNo
- Trains on your dataNo, but data in cloud
- Local supportNo
- Pricing modelToken-based
- Suited for critical infraLimited
QWorker
Local & EU
- Where does the AI run?Local / on-prem / EU
- Data residencyEU — you choose
- GDPR & AI ActDPA, ISAE 3000, AI Act-ready
- Remembers your contextYes — local memory layer
- Multi-model routingYes — you approve the models
- Acts in your systemsReady-made + custom EU connectors
- Agents per departmentBuilt for your processes
- Air-gapped optionYes
- Trains on your dataNever — data stays on-site
- Local supportYes, 24/7
- Pricing modelFixed price per user in EUR
- Suited for critical infraYes — NIS2, DORA
The autonomous analyst
QWorker works while the organization sleeps — indexing documents, summarizing changes and updating knowledge bases.
Nightly indexing
Automatic local summaries of changes in selected folders and workflows.
Historical context
Reuse prior analyses, answers, templates and projects actively.
Version control
Rewind and understand exactly what changed and when.
- 22:00Indexes new documents
Picks up every file changed during the day in the folders you selected.
- 01:30Summarises and links
Generates summaries, extracts action items and links related cases.
- 04:00Updates the knowledge base
Writes the results into your internal wiki, CRM or case system.
- 07:30Briefing is ready
Your team starts the day with a fresh briefing — no manual prep.
How a document is processed
The whole flow stays inside your perimeter. Nothing leaves the network — not the prompt, not the context, not the answer.
Document
PDF, Word, email, code
Local AI engine
Runs in your environment
Your data
Context fetched locally
Output
Back to the user
- · Documents and context
- · Prompts and intermediate results
- · Model weights and embeddings
- · Audit log and user activity
- · OpenAI, Anthropic, Google and similar
- · US hyperscalers
- · Telemetry containing content
- · Training data shared with third parties
Full data security via local or air-gapped deployment
QWorker runs on your own hardware, integrates with Active Directory and is segmented by role or clearance level.
No public cloud dependency
Run locally with open-source or controlled models.
AD-based access control
Departments and users only see the data they're allowed to see.
Closed loop
Data never leaves your network or your hardware.
Concrete commitments — not marketing
Local execution
Runs on your own hardware or private cloud — never on public models.
Zero data leakage
Prompts, documents and output never leave your network.
Air-gapped & controlled
Supports fully isolated environments with segmented access and audit logs.
What does that mean for you?
- Sensitive documents can be analyzed without NDA risk
- Compliance with GDPR, NIS2 and internal data policies
- No dependency on US cloud providers
- Full audit trail of who asked what
- AI can be used in classified and closed environments
All AI calls are hosted in the EU — documented per provider and region
QWorker is built on one principle: your data never leaves Europe. We use only AI providers with EU hosting, zero retention and contracts under European jurisdiction — so you can show regulators, customers and the board exactly where data lives.
EU AI gateway
Sweden Central (SE)Traffic stays within the EU data boundary — no data leaves for inference or logging.
Mistral AI
Paris (FR)European provider hosted in France under French and European jurisdiction.
QWorker Cloud (database & files)
Frankfurt (DE)All persisted data — projects, todos, chats, files — sits in the EU and is encrypted at-rest.
Transactional email
EU regionInvitation and notification emails are sent from EU infrastructure with no export to third countries.
How data is handled — without leaks
Every prompt follows a mapped route. No stopovers outside the EU, no third party allowed to train on the content, and no shadow copies in US cloud regions.
Input — Browser → QWorker Edge
EU edge — Frankfurt / Amsterdam / StockholmPrompten sendes fra din browser over TLS direkte til en QWorker edge-node i EU. Ingen mellemstation uden for Europa.
Prompt-tekst, eventuelle vedhæftninger, thread-id og brugerens JWT.
Browser → QWorker Edge (Cloudflare Workers)
- TLS 1.3 mellem browser og edge
- JWT bearer verificeres pr. request
- Content-Security-Policy blokerer 3.-parts scripts
- Cloudflare-routing låst til EU-regioner
- Ingen logging af prompt-indhold på edge
- Ingen US-region fallback ved load
No training on your data
Our AI providers do not train models on prompts or responses from QWorker. Your input is used only to generate the current response.
Zero retention on inference
Prompts are not stored by the AI provider beyond the short window required to return an answer. What we log, we log ourselves — in the EU.
Encryption in transit and at rest
TLS 1.2+ between all hops and AES-256 on the database. Access tokens are rotated and service keys stay out of the client bundle.
RLS on every table
Row Level Security in the database ensures users only see their own projects, todos, meetings and files — even if an API call were misused.
DPA ready to sign
DPA with sub-processor list (Microsoft Ireland, Mistral) is pre-filled and ready for your DPO to sign.
No US transfers
We do not use OpenAI directly, Anthropic, Google Gemini or other US-hosted models. Standard Contractual Clauses are not needed — everything stays in the EU.
Want the documentation in writing? We'll send the data card, sub-processor list and DPA by email.
What compliance teams usually ask
Short, concrete answers about hosting, retention, logging and access control — so you can share them directly with your DPO.
AI Act, GDPR and AD — documented in one connected track
QWorker is built for European organisations that need to prove compliance — not just claim it. We collect legal requirements, access control and documentation in one place, so leadership, DPO and auditor see the same picture.
AI Act compliance from day one
If your employees use ChatGPT, Copilot or HR software with AI, you are legally a "deployer" under the AI Act. That triggers concrete requirements which QWorker helps you meet and document.
AI literacy (article 4)
Built-in AI awareness material and "AI license" flow for employees — with receipt and logging so training can be documented at inspection.
Risk classification
Assessment of your AI use against the AI Act's risk categories (limited, high, unacceptable) and automatic flagging of use cases that need extra controls.
Use register & transparency
Every agent has a register of purpose, data, model and human oversight — ready for documentation to leadership, DPO and regulator.
AI policy as a living document
A template for your AI policy, linked to actual use in QWorker, so policy and practice always match.
GDPR in practice — not only on paper
The classic mistake is confidential company or customer data ending up in an open AI model. QWorker keeps data inside your perimeter and delivers the paper trail GDPR requires.
Data minimisation & local processing
Models run locally or on-prem. Personal data never leaves your infrastructure — not even for third-party training.
Processing register
Automatic Article 30 register per agent: purpose, categories, recipients, retention and security measures.
DPIA template
Pre-filled impact assessments (Article 35) for the most common AI use cases — ready for your DPO's review.
Incident log & breach handling
Immutable log of accesses and prompts, so you can react within the 72-hour window on a breach — and prove nothing was leaked.
AD-based access control
QWorker mirrors your existing role model in Active Directory, so the AI only sees what the user already has access to — and you avoid shadow AI outside your entitlement structure.
SSO via AD / Entra ID
Single sign-on and automatic provisioning. When an employee leaves, AI access closes with the AD account.
Role and attribute based access
Every agent and dataset is bound to AD groups and attributes — from "HR manager" to "classified up to secret".
Segmentation & need-to-know
The AI respects NTFS and SharePoint permissions. A salesperson never sees HR data, and a consultant never sees another customer's case.
Audit trail per user
Every prompt, response and action is logged against AD identity — hooks straight into your SIEM or Microsoft Purview.
What you receive — ready for auditor and regulator
Instead of gathering compliance evidence from scattered systems, QWorker generates the documentation continuously. When the regulator, a customer or your board asks, the answer is a few clicks away.
Get the compliance package reviewed- Processing register (GDPR art. 30) per agent — export to Excel/PDF
- AI use register with purpose, model, data sources and human oversight
- Pre-filled DPIA template for the most common use cases
- AI policy template linked to actual QWorker use
- AI literacy documentation: participants, content and receipts
- Immutable audit log of prompts, responses and accesses per AD user
- Risk classification per agent against the AI Act categories
- Vendor declaration and DPA ready for your compliance team
Concrete numbers, not loose promises
Every criterion is measurable, documented and delivered as part of the rollout.
Encryption
AES-256 at rest · TLS 1.3 in transitAll documents, embeddings and logs are encrypted on disk and over the network.
Keys live in your KMS/HSM — we never have access.
Access control
RBAC + SSO (SAML, OIDC) · MFA requiredRole-based access down to document and feature level, integrated with your IdP.
Ready-made roles for admin, user and auditor — extensible as needed.
Audit log
Append-only · 100% of requests · 12-month retentionEvery prompt, source and output is logged immutably with user and timestamp.
Export to SIEM (Splunk, Sentinel, Elastic) out of the box.
Key management
BYOK · rotation every 90 daysYou own and rotate every encryption key. No shared secrets.
Integrates with Azure Key Vault, AWS KMS or HashiCorp Vault.
Data isolation
Single-tenant · network segmentedNo resource sharing with other customers. Models and data live in your VPC or on-prem.
Dedicated instance with documented network topology.
Monitoring & SLO
99.9% uptime · alerts in under 5 minReal-time metrics on latency, error rates and anomaly detection on usage patterns.
Dashboard and webhook alerts to Teams, Slack or PagerDuty.
Vulnerability management
CVE patching < 7 days · quarterly pentestContinuous dependency scanning and a documented remediation process.
Report after every pentest plus access to the SBOM.
Model governance
Versioned models · signed artifactsEvery model update is traceable, approved and reversible.
Changelog and approval flow with your compliance team.
See QWorker run inside a closed environment
A realistic walk-through from a government case — no internet, no external cloud, no data sharing.
Classified material is loaded
A case worker adds 240 documents from an internal case — everything happens inside the closed network.
Context is fetched locally
QWorker indexes and finds relevant passages with embeddings computed on-prem.
Draft is generated
A memo is assembled with citations to the original documents — no hallucinations without a reference.
User reviews and signs off
Output is reviewed, edited and recorded in the audit trail. Nothing leaves the environment.
What does it take to run this in your environment?
1 server with 2× NVIDIA L40S or equivalent
Scales from 20 to 2,000 users per node.
Linux (RHEL, Ubuntu LTS) · Docker or Kubernetes
Delivered as a signed offline bundle.
No outbound internet required
Updates delivered through an approved media channel.
Min. 2 TB encrypted disk for models and indexes
Backup integrates with your existing routines.
SAML/OIDC against your internal IdP · MFA
Works with Active Directory, Keycloak and others.
Signed model artifacts and SBOM
Ready for verification by your security team.
Which setup fits you?
Four short questions — and we'll suggest a concrete demo that matches your security needs.
How sensitive is the data your AI will work with?
From quick validation to critical infrastructure
Consulting firms and SMBs
Easy install, fast value and low friction.
Enterprise teams
Integration with internal systems, AD and team-specific knowledge.
Government and defense
Local, security-cleared AI for environments where cloud isn't an option.
AI should move into the workflow, not out into the browser
The next wave of AI isn't about better prompts — it's about AI that already understands the context, works close to the data and respects your security requirements.
Understand your responsibility and the new legal requirements as an employer
Many companies mistakenly think the AI Act only applies to tech giants. But if your employees use ChatGPT or Copilot, or if you use software for recruiting and scheduling, you are legally a "deployer" — and covered by concrete requirements that apply already now.
The legal AI-literacy requirement
It is now a legal requirement that employers train employees in safe use of AI. We show you what the law requires and how to easily give employees the necessary "AI license".
Data security & GDPR in practice
How to avoid the classic (and expensive) mistake of confidential company or customer data ending up in an open AI model.
Practical AI awareness material
A concrete AI awareness example you can take with you and use in your organisation right away.
A clear AI policy
Your most important tool to document compliance and create trust for both management and employees — so you can leverage AI effectively, safely and fully lawfully.
We boil the law down to practical reality, so you can act right away — without legal detours.
Frequently asked questions
Start for free. Scale when you're ready.
Transparent pricing in euro. No hidden usage fees for tokens or compute — your data and your budget stay in the EU.
Pilot
30-day free pilot
Try QWorker in your own environment — no commitment.
Included in Pilot
- Up to 10 users
- Runs locally on your infrastructure
- Ready-made connectors for Microsoft 365 & Google
- Standard security model (RBAC, audit log)
- Onboarding workshop (2 hours)
- Community support
Team
per user / month — billed yearly
For departments that want to automate daily work.
Everything in Pilot, plus
- Unlimited users
- You choose and approve the models
- Memory and personalisation per team
- Workflows and agents across systems
- SSO (Entra ID, Google Workspace)
- Priority support
Business
per user / month — from 50 users, yearly
For organisations with strict operational and security needs.
Everything in Team, plus
- Dedicated instance in the EU or with you
- DPA, ISAE 3000 and full GDPR documentation
- Advanced role and permission management
- SIEM integration and extended audit
- Custom connectors and API access
- Named customer manager
Enterprise
Air-gapped, on-prem or sovereign cloud
For critical infrastructure, public sector and regulated industries.
Everything in Business, plus
- Air-gapped and on-prem deployment
- Your own LLM models, fine-tuning and hosting
- NIS2, DORA and sector-specific compliance
- Penetration testing and customer-specific threat model
- SLA with local 24/7 support
- Custom contract and billing
All prices ex. VAT. Non-profit and education: ask about our discount.
Book a meeting
30 minutes with QWorker
Pick a time that suits you. We'll walk through your setup and show concretely how local AI can solve your tasks — without sending data out of the house.
No available times that day — pick another date.
Developed by q-analytics
QWorker is designed and built by q-analytics — a Danish data and AI consultancy specialised in secure, locally-deployed AI for regulated industries. We own the product, the engineering and the security architecture end-to-end.
Read more about the partnership