Can the system use our documents?
Project sources can be organized through Sanskrit so a supported workflow retrieves passages from authorized material.
LodzIA / Private AI
Ask questions about your documents, search the web when needed, and get answers with links to their sources.
Open workspace ↗01 / 03
Find the information you need in your documents. Follow the links in the answer to read the original material and check the details.
02 / 03
Identity and authorization determine who can reach a workspace and its resources. Aegis carries the operation lifecycle and telemetry across the request.
When a project is backed by Sanskrit, source files are extracted and organized into retrievable passages. Relevant passages retain a link to the source material.
When you choose web research, Ethicrawler finds relevant pages and keeps their links so you can read the original sources.
Gufé processes the model request on server GPUs in a private data center or controlled cloud. Model inputs are not sent to an external AI provider.
Read the answer and open the documents or web pages used to prepare it.
03 / 03
The browser is the interface; inference runs on a dedicated GPU server. Deployment can be in a private data center or controlled cloud.
Project sources can be organized through Sanskrit so a supported workflow retrieves passages from authorized material.
No. Web research is an optional workflow and depends on configured services and user action.
Aegis provides operation and telemetry foundations. The accessible views depend on role and deployment configuration.
LodzIA exposes an API with managed access keys. Available capabilities follow the services configured for the deployment.
Open the workspace, add a few documents and ask a question. The models run on dedicated GPUs, and answers based on your documents show where each passage came from.
Open workspace ↗