Local AI Station with integrated RAG: your documents searchable out of the box
Imagine an AI that already knows your documents. You ask a question about a contract, a report, or an internal procedure, and it answers by citing your own files, all while keeping your data within your office. This is what RAG enables. On our workstations, it comes ready to use: zero configuration, with your documents ready to be queried the moment you unbox.
The problem RAG solves
No matter how powerful an AI model is, it does not know your specific documents. It was trained on general data, not on your contracts, files, or internal documentation.
The result: if you ask it a specific question about your business, it will improvise or make things up. This is the well-known "hallucination." It is unusable for professional work.
Without RAG
"What does clause 4 of my standard contract say?" The AI has never seen it. It provides a vague answer or invents a plausible but incorrect clause.
With RAG
The same question. The AI finds your contract, reads clause 4, and answers by quoting the exact text. Reliable, verifiable, and based on your documents.
RAG explained simply
RAG stands for "Retrieval-Augmented Generation." The concept is straightforward: instead of relying solely on the model's internal memory, it is given access to your specific documents at the moment it generates a response.
In practice, your files are broken down, indexed, and stored in a specialized database. When you ask a question, the system retrieves the relevant passages and provides them to the model, which then crafts a response based on your actual data.
Upload
PDF, Word, Excel, text: your documents are imported onto the machine.
Indexing
Documents are split and indexed locally, only once.
Query
Ask your questions in natural language, and the AI responds while citing your files.
Why "zero configuration" is a game-changer
Setting up a RAG system yourself is no simple task. You have to select an embedding model, choose a vector database, fine-tune document chunking, and then connect everything to a language model and a user interface. Every step introduces a point of friction and a potential source of silent errors.
For professionals outside the IT field, this represents a major hurdle. Many local AI projects stall at this very stage.
Who is a local RAG for?
- Lawyers and notaries: query a database of contracts, case law, and templates while maintaining professional privilege.
- Doctors: find information within protocols and recommendations without exposing patient data.
- Chartered Accountants: query doctrine, internal procedures, and files while keeping everything on-site.
- Design offices and engineers: search through extensive standards, specifications, and technical reports.
- Businesses: provide your teams with an AI that understands internal documentation, procedures, and history.
- Researchers: query a library of articles and notes using natural language.
Optional RAG on all our AI stations
The RAG option is available on all of our AI stations, from the CoreIA RTX 5060 Ti to our rack servers. You can add it in two ways: directly via the online configurator on the product page, or through a custom quote if you have specific requirements (large volumes of documentation, specific formats, multiple separate databases, or multi-user access). RAG can be integrated into any local AI station in our range.
Describe your documents and your intended use, and we will configure the appropriate RAG before shipping. Contact us at contact@radiancesystems.eu or use the quote form on our website.
What the RAG option includes
- A language model tailored to your needs, pre-installed and ready to use.
- A local indexing engine to transform your documents into a searchable database.
- A documentation database that remains entirely on your machine.
- A chat interface accessible via a browser, similar to ChatGPT.
- Support for common formats: PDF, Word, Excel, and plain text.
- Easy setup to upload your documents and start asking your first questions.
Which workstation for a local RAG?
RAG works on all our workstations. Your choice depends primarily on the volume of documents and the number of users. The larger the database and the more demanding the models, the more critical VRAM and RAM become.
CoreIA RTX 5060 TiIndividual RAG, standard document databases.
€2,068
CoreIA Radeon AI PRO R9700RAG for large volumes, 30B models in 8-bit.
€4,162
CoreIA RTX 5090RAG with 30B models in 8-bit, high-quality responses.
€10,291
CoreIA GB10 (ASUS Ascent GX10)Shared RAG for small teams via the local network.
€6,999
For large enterprises and multi-user access
When RAG must serve numerous employees simultaneously across large document databases, a high-end server is essential. These configurations provide the VRAM, power, and reliability needed to run large models and serve multiple users at once without any slowdown.
CoreIA 2x RTX 5090 (4U rack, 64 GB)Multi-user RAG server, high volume, 4U rack format.
€20,954
CoreIA Threadripper PRO 4 GPU96 GB of VRAM (up to 4 GPUs), ECC RAM, massive document databases, enterprise-grade reliability.
€32,647
CoreIA 2x RTX PRO 6000 (192 GB ECC)Enterprise-scale RAG, larger models, 24/7 availability.
€48,508
Frequently Asked Questions
What is a local AI workstation with integrated RAG?
This machine runs AI locally and comes with a pre-installed and configured RAG system. Simply upload your documents, and the AI can answer your questions in natural language, requiring no technical configuration and ensuring no data ever leaves your machine.
Is RAG included by default?
RAG is available as an option on all our AI workstations. You can add it through our online configurator on the product page, or via a custom quote for specific requirements. We then configure the entire chain before shipping.
Which document formats are supported?
Common formats include PDF, Word, Excel, and plain text. You can import contracts, reports, procedures, and knowledge bases. The system indexes them locally once, allowing you to query them at any time.
Do my documents remain confidential?
Yes, completely. With a local RAG, both indexing and querying take place on your machine. No documents are sent to an external server. This makes our solution ideal for professionals bound by confidentiality agreements or GDPR regulations.
Do I need technical skills to use it?
No. With the ready-to-use RAG option, everything is pre-installed. The interface functions like a standard chat tool. You simply add your documents and ask your questions without ever needing to use a command line.
Can I update the document database?
Yes. You can add new documents at any time; they will be indexed and become searchable immediately. The database grows along with your business.
Which machine should I choose for a local RAG?
For individual use with a standard database, the CoreIA RTX 5060 Ti (starting from €2,068) is suitable. For high volumes or shared team access, the CoreIA Radeon AI PRO R9700, the CoreIA RTX 5090, or the CoreIA GB10 are better suited. RAG remains available as an option on every model.
