UN-INTELS
Local AI & Microsoft Office Integration Bridge
Connect your offline local AI models running in LM Studio directly with Microsoft Excel, Word, and PowerPoint. Zero cloud telemetry, pure Windows COM automation, and full Administrator execution authority.
Connecting to active Microsoft Excel instance...
[v] Successfully hooked into active Excel!
[+] Connected to Local AI Server (http://localhost:1234/v1). Active Model: openai/gpt-oss-20b
Un-Intels [Excel] > do a simple example salary sheet of 10 peoples
[Processing request via Local AI...]
Executing action on Microsoft Excel...
[v] Action executed successfully on Excel! (10 Rows, Formulas & Currency Formatted)
NECESSARY TOOLS & DOWNLOADS
Open_Script.bat
Self-elevating Administrator launcher. Resolves system Python paths and executes un_intels.py cleanly.
un_intels.py
Main Python automation bridge engine featuring Win32 COM hooks, LM Studio API client, and self-healing execution.
LM Studio
Required local LLM runner. Download LLM models (e.g. GPT-OSS, Llama 3, Qwen) and start the local API server on port 1234.
Python 3.10+
Required Windows Python environment. Ensure pywin32, xlwings, and requests libraries are installed.
STEP-BY-STEP TRACK TUTORIAL
Install Prerequisites & Dependencies
Ensure Python 3.10+ is installed on your Windows system and added to your system PATH. Install the required Python libraries using pip:
pip install pywin32 xlwings requests psutil
Download & Place Un-Intels Files
Download Open_Script.bat and un_intels.py into your target project directory (e.g. X:\NaroHostBase\Un-Intels\).
Launch Local AI Server in LM Studio
Open LM Studio, download a chat model (e.g. openai/gpt-oss-20b or llama-3), and enable the Local API Server on port 1234 with CORS enabled. Alternatively, Un-Intels can auto-launch the server when started.
Execute Open_Script.bat with Admin Privileges
Double-click Open_Script.bat and select Yes on the Windows UAC Prompt. The script will hook into your active Microsoft Excel, Word, or PowerPoint application.
Interact & Enable Auto-Execution Mode
Type your request in plain English (e.g. do a simple example salary sheet). Type /a to toggle Auto-Execution Mode, allowing Un-Intels to write formulas, tables, and formatting directly onto your active sheet without confirmation prompts.
SHORTCUT COMMAND DIRECTORY
| COMMAND | SHORTCUT | FUNCTION & DESCRIPTION |
|---|---|---|
| /auto | /a | Toggle direct Auto-Execution mode (skips Y/N confirmation step). |
| /models | /m | Scan disk models and load selected LLM chat model into LM Studio memory. |
| /switch | /s | Switch active target Office application (Excel / Word / PowerPoint). |
| /open | /o | Restore and bring LM Studio & Microsoft Office GUI windows to front. |
| /status | /st | Display full system status dashboard and API connection metrics. |
| /reset | /r | Reset current conversation memory and system prompt context. |
| /url | /u | Change or update local AI endpoint URL (default: http://localhost:1234/v1). |
| /cls | /c | Clear terminal screen and re-render header control box. |
| /exit | /q | Quit and safely exit Un-Intels application. |