KNOWLEDGE BASE
Upload company documents and make your AI answer questions from them — product manuals, legal docs, SOPs, financial reports, anything.
How it works
When you upload a document, NestAI sends it to your private server where it's split into chunks and converted to vector embeddings using the Nomic Embed model. When a user asks a question in Open WebUI, the system finds the most relevant chunks from your documents and includes them as context for the AI — this is called RAG (Retrieval Augmented Generation).
nomic-embed-text model must be installed on your server for the knowledge base to work. Install it from Dashboard → Models if it's not there.Uploading documents
Go to Dashboard → Knowledge. Drop a file on the upload zone or click to browse. Supported formats:
| Format | Extension | Notes |
|---|---|---|
| Best format — preserves structure | ||
| Word Document | .docx | Tables and lists supported |
| Plain text | .txt | Simple and fast to process |
| Markdown | .md | Good for technical docs |
| CSV | .csv | For data and tables |
Max file size is 50MB. Processing takes 30 seconds to a few minutes depending on document length.
Using your knowledge base in chats
After uploading, enable the knowledge collection in Open WebUI:
- 1Open your AI at yourname.nestai.chirai.dev
- 2In the chat input, click the + icon or the document icon
- 3Select your knowledge collection from the list
- 4Ask questions — the AI will reference your documents in its answers
Deleting documents
Click ✕ next to any document in the Knowledge Base page. This removes it from both NestAI and your server's vector store. Existing conversations are not affected but the AI will no longer have access to that content.
Limits
| Limit | Value |
|---|---|
| Max file size | 50 MB per file |
| File formats | PDF, DOCX, TXT, MD, CSV |
| Number of files | No hard limit — constrained by server disk space |
| Processing time | 30 sec – 5 min depending on size |