- Home
- Skills
- Agents & Orchestration
- Retro AI
Works with the AI Agents you already use
Retro AI
A fully local, privacy-first AI assistant that runs on your machine via Ollama — no cloud, no tracking, no API keys.
$3
Retro AI
Example session with this skill installed
/search latest Python 3.12 features then create a folder named 'Python_Update' on my desktop and save the summary as a text file there.
- Read your context and instructions
- Compiled the retro ai
I have performed the web search, synthesized the key features of Python 3.12, and executed the file system commands on your Windows machine.
Local Automation Summary
| Action | Status | Details |
|---|---|---|
| Web Search | Success | Fetched via DuckDuckGo/Wikipedia |
| Directory Creation | Success | C:\Users\Admin\Desktop\Python_Update |
| File Write | Success | Created features_summary.txt |
| System Intent | Complete | Multi-intent chain finished |
Python 3.12 Key Features Extracted
F-string improvements: Removal of many previous restrictions, allowing nested f-strings and multiline expressions.
- TypedDict enhancements: More precise type hinting for dictionary structures.
- Performance: Significant speed improvements in the interpreter and specialized type variables.
- Error Messages: More descriptive suggestions for common NameErrors and SyntaxErrors.
Next steps
- Use
/fetchon specific PEP documentation URLs for deeper technical analysis. - Ask me to zip the 'Python_Update' folder if you need to share the research.
- Resume this session later using
python retro.py --session python_research.
Connects securely to your tools. The creator never sees your data.
About this skill
Cloud-based AI assistants require sending sensitive data, files, and queries to external servers. Retro AI eliminates this privacy risk by running a powerful agentic assistant entirely on your local Windows hardware.
What it does
- Persistent Memory provides cross-session conversation history using a local SQLite database to maintain context.
- System Control manages Windows operations including launching apps, file manipulation, and volume or brightness adjustment.
- Real-Time Data fetches web searches, weather, stock prices, and news headlines without requiring external API keys.
- Screen Vision captures and analyzes your desktop using local vision models like qwen3-vl or llava.
- Automation Engine executes Python code, shell commands, and browser automation via Playwright or pyautogui.
How it works
- Launch the assistant via
python retro.pyto enable the local memory and tool execution layer. - Provide natural language commands for system tasks or use slash commands like
/searchor/stockfor real-time data. - Use the
--sessionflag to resume specific projects with their full historical context. - Chain multiple intents in one message to trigger sequential automation workflows.
Frameworks & tools
Built for Windows 10/11 using Python 3.10+, Ollama, and Qwen3 8B. It utilizes SQLite for storage and optional libraries like Playwright, pyautogui, and Pillow for advanced automation.
Why this beats prompting it yourself
Standard Ollama models are stateless and lack system access. Retro AI adds the necessary orchestration layer to bridge the model to your local files, system settings, and persistent database.
Use cases
- Automate repetitive Windows file management and application workflows locally.
- Conduct private research using local web search tools and persistent session memory.
- Debug and execute Python scripts directly through a conversational interface.
- Analyze sensitive on-screen data using local vision models without cloud uploads.
Known limitations
Requires Windows 10/11 for system control features. Performance depends on local GPU VRAM, optimized for cards with 8GB or more. Internet is required for real-time data tools.
Requires Windows 10/11 for system control. Performance scales with local GPU VRAM. Internet access needed only for real-time data tools. AI inference is strictly local.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 8 days ago
Requires
Permissions
File Scopes
Network access is only used for real-time data tools (weather, news, stocks, web search). All AI inference runs locally via Ollama. No data is sent to any cloud service.