Local AI assistant

AI chatbot that answers from your own documentation

A private AI assistant for onboarding, support and internal knowledge access — running on your server or EU hosting, without sending sensitive data to public AI platforms.

Request a free consultation → View FAQ

✅ Runs on your infrastructure ✅ Answers from your documentation ✅ No OpenAI per-message costs ✅ Website widget or internal assistant

What problem does this solve?

⏳

Employees keep asking the same questions because knowledge is not easy to find.

📚

Your documentation exists, but people do not use it because searching through it takes too long.

🎧

Support or internal teams spend too much time answering repetitive questions.

🔒

You want AI assistance, but you do not want company data sent to public third-party systems.

This is not a generic chatbot. It is an AI assistant connected to your own documentation, processes and internal knowledge.

Who is this for?

Internal help desk and IT support

For teams that want faster answers to internal questions about setup, access, procedures and common issues.

Onboarding and internal knowledge

For companies that want new employees to get answers from structured documentation instead of depending on one person.

Customer support and website assistant

For businesses that want a chatbot widget on their site that answers based on approved content and documentation.

Privacy-conscious organizations

For teams that want local AI capabilities without exposing data unnecessarily to public AI services.

How it works

Your documentation

We connect your guides, procedures, FAQ pages and internal knowledge base.

Search and retrieval

The system finds the most relevant parts of your documentation for each question.

Local AI response

The assistant generates an answer using your own knowledge base, not random internet content.

Continuous updates

When your documentation changes, the assistant can be updated automatically so answers stay current.

What is needed for it to work well?

✅ Existing documentation or FAQ content

✅ A Linux server or EU-hosted environment

✅ Clear internal procedures or knowledge sources

✅ Questions you want the assistant to handle

✅ A realistic scope for the first version

✅ Willingness to improve documentation over time

Typical use cases

Employee onboarding

New team members can ask questions about procedures, tools and internal rules without waiting for manual answers.

Internal IT support

The assistant answers frequent questions about VPN, passwords, access requests, software setup and internal instructions.

Website support widget

A chatbot widget can answer visitors using your approved business content, FAQ or documentation in real time.

Local AI assistant vs public AI service

More control

Your data and knowledge stay closer to your own environment instead of being unnecessarily exposed to external platforms.

Predictable cost model

Instead of paying per message or token usage, you can run the system on your own infrastructure.

Documentation-based answers

The assistant responds from your actual content instead of relying on vague or generic responses.

Business-specific setup

The solution is adapted to your internal knowledge, language and workflow instead of being a one-size-fits-all chatbot.

Technology stack

The implementation uses a practical open-source stack designed for private AI assistance and documentation-based responses.

BookStack Knowledge base and documentation source

n8n Automation and content update workflows

Weaviate Vector database for semantic search

Ollama Local model runtime

RAG architecture Answers grounded in your documentation

Website widget Optional public-facing chatbot integration

Frequently asked questions

How long does setup take? ▾

If you already have a Linux server and documentation, setup typically takes 2 working days. This includes the full stack installation, importing your documentation and testing with your own questions.

What server is required? ▾

RockyLinux or AlmaLinux 8+ is recommended, with at least 8 GB RAM and 2 vCPU. For more serious use, 8 GB RAM is the practical minimum.

What if the assistant gives a wrong answer? ▾

With a RAG-based setup, the assistant answers from your documentation. If the answer is missing or wrong, it usually means the source documentation needs improvement or clarification.

Can it be added as a website widget? ▾

Yes. A chatbot widget can be integrated into WordPress or another website so visitors can ask questions and get responses in real time.

How is documentation updated? ▾

Documentation is updated directly in BookStack. Once content is saved, the workflow can automatically refresh the vector database so the assistant sees changes within minutes.

Which LLM model is used? ▾

We use Ollama to run open-source models such as Llama, Mistral or Gemma, depending on your hardware capacity and language requirements.

Want a private AI assistant for your business?

Tell us what documentation you already have and what kind of questions you want the assistant to handle.

Request a free consultation →