- LOCAL
- RELEVANT
- SECURE
- HUMAN
Short answer
How can I chat with my private files without uploading them?
You use local embeddings to turn your documents into searchable math that stays on your machine. Your AI client then uses reranking to pull only the most relevant snippets for its response. This setup ensures your raw data never touches a cloud server.
The results.
Where your data goes.
The process moves from raw files to precise answers through local computation.
LOCAL
Data stays home
The MCP processes your files on your own hardware. No sensitive text is sent to external training sets or cloud storage.
RELEVANT
Precise context retrieval
The system finds the exact paragraphs needed for your question. This prevents the AI from getting lost in massive documents.
SECURE
Privacy by design
Since the embeddings are generated locally, your document structure and content remain under your control.
HUMAN
You verify everything
The AI provides answers based on the retrieved text. You decide if the logic holds up and whether to act on the information.
The workflow
What your AI does when the files are ready.
The AI follows a specific path to turn your local files into usable knowledge.
Index your files
The system converts your local text into vector representations so they can be searched mathematically.
create_embeddingsFind the right context
When you ask a question, the system identifies the most important sections of your documents.
rerank_documentsGenerate the answer
The AI uses the filtered, highly relevant snippets to construct a direct response to your query.
chat_completionsRefine the search
If the first pass misses the mark, the system can re-evaluate the document chunks to find better matches.
rerank_documentscreate_embeddings
Try it
Copy these to start.
Use these prompts once you have connected LocalAI.
Starting points
These are opening lines. Replace the bracketed text with your specific file names or folder paths.
LocalAI Connector
You're all set. Choose your MCP client and follow the setup instructions.
https://edge.vinkius.com/vk_preview_rNQOAeEr0Jny5r5Ag0JD5dh4JLxiMfgFSBUDjX1k/mcpClaude Desktop
Follow the steps below to connect in seconds.
- 1In Claude Desktop, open Settings → Connectors.
- 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
- 3Click Add and start a new chat — LocalAI capabilities are ready to use.
{
"mcpServers": {
"localai-mcp": {
"url": "https://edge.vinkius.com/vk_preview_rNQOAeEr0Jny5r5Ag0JD5dh4JLxiMfgFSBUDjX1k/mcp"
}
}
}Summarize the main points in the legal_contract.pdf file.
Based on the project_notes.txt, what are the three biggest risks identified?
Find the specific clause in the employee_handbook.pdf regarding remote work.
Compare the budget figures in the Q3_report.csv with the Q4_report.csv.
Claude
ChatGPT
Cursor
VS Code
Windsurf
Claude Code
JetBrains
Cline
Start here
Connect LocalAI once, then ask.
Click the link to connect the MCP to your client. Your credentials stay encrypted, and you can start querying your local files immediately.
Connect LocalAI to your AIFAQ
How this task behaves
- 01
Does my data get sent to the cloud for training?
No. All embedding and reranking happens locally on your machine.
- 02
Can the AI edit my original documents?
No. This MCP only reads and analyzes your files. It cannot change or delete your data.
- 03
How does the AI know which part of the file to read?
It uses local embeddings to find relevant sections and then reranks them to ensure the most useful context is sent to the chat.
- 04
Do I need to upload my files to Vinkius?
No. Vinkius hosts the connection, but the actual data processing stays local via the MCP.
- 05
Can the AI write new files to my hard drive?
No. The available tools are limited to creating embeddings, reranking, and generating chat completions.
More questions about LocalAI? The Connector page answers them. See everything the LocalAI Connector can do
Connect LocalAI to Claude, Cursor, ChatGPT & more
