- VECTOR
- RETRIEVAL
- KNOWLEDGE
- HUMAN
Short answer
How do I make my AI search by meaning instead of keywords?
You use this MCP to turn your text files into vector embeddings. This lets your agent find the right context by looking for semantic similarity. You'll see how the process works below.
Search outcomes
What your data becomes
The results of processing your files through the agent.
VECTOR
Mathematical maps
The AI turns text into high-dimensional vectors. These numbers represent the core meaning of your documents.
RETRIEVAL
Contextual hits
When you ask a question, the agent finds the closest matching vectors. This pulls up relevant info even if the words don't match exactly.
KNOWLEDGE
Deep understanding
Your agent gains a better grasp of your specific files. It can now answer questions based on the nuances in your data.
HUMAN
Manual control
You decide which files to index. The AI won't change or delete your original files without your direct command.
The workflow
What your AI does when the data arrives.
The agent handles the heavy lifting of converting and organizing your text.
Find files
The agent scans your directory to identify the text files you want to index.
list_filesCreate vectors
The agent sends the text to OpenAI to generate a numerical representation of the content.
create_embeddingMap meanings
The agent uses the resulting embeddings to build a searchable map of your information.
create_embeddingVerify contents
The agent checks the file list to ensure all intended documents are accounted for.
list_files
Try it
Copy these to start.
Use these prompts to kick off the process.
Starting points
Swap the bracketed text for your actual file names or folder paths.
OpenAI Connector
You're all set. Choose your MCP client and follow the setup instructions.
https://edge.vinkius.com/vk_preview_wqDLmxGMRbmUrKtHE8ZWJc73FuzFzNT70ItXxJkE/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 — OpenAI capabilities are ready to use.
{
"mcpServers": {
"openai-mcp": {
"url": "https://edge.vinkius.com/vk_preview_wqDLmxGMRbmUrKtHE8ZWJc73FuzFzNT70ItXxJkE/mcp"
}
}
}List all the files in my current directory so we can see what to index.
Create embeddings for the project_notes.txt file.
Look through my documentation folder and create embeddings for every text file you find.
Find the most relevant parts of the user_manual.txt based on this question: [insert question].
Claude
ChatGPT
Cursor
VS Code
Windsurf
Claude Code
JetBrains
Cline
Start here
Connect OpenAI once, then ask.
Just link your OpenAI account. Your credentials stay encrypted, and you can start using these tools in your favorite MCP client immediately.
Connect OpenAI to your AIFAQ
How this task behaves
- 01
Can the AI delete my original text files?
No. The tools only allow the AI to list files and create embeddings. It cannot modify or delete your data.
- 02
Does this work with PDFs?
No. Currently, the tools are designed to work with text-based data files.
- 03
Will the AI store my embeddings locally?
The AI uses the embeddings to perform searches within your session, but this MCP does not manage a permanent local database for you.
- 04
Can I use this to search my entire hard drive?
No. You must point the agent to specific files or directories using the provided tools.
- 05
Does the AI see my OpenAI API key?
No. Your connection is managed securely through Vinkius; your agent only interacts with the tools provided.
More questions about OpenAI? The Connector page answers them. See everything the OpenAI Connector can do
Connect OpenAI to Claude, Cursor, ChatGPT & more
