The Backstory: Engineering "Domestic Bliss"
As an OS developer and hardware entrepreneur with over 20 years in the game, my life is usually defined by "Walking & Coding." However, even the most robust system architecture can’t solve a simple domestic pain point: "Honey, can you find this movie for me?"
The old workflow was tedious: Get the request -> Manual search for magnets -> Log into the NAS -> Paste the link -> Monitor the download. It felt like a waste of cycles.
Since I'm currently building Hermes (an AI Agent) and the LCMD private cloud hardware, I decided to close the loop. I wrote a Skill that allows the Agent to handle the entire "Search-to-Download" pipeline via natural language.
The Stack: AI Agent + NAS + VueTorrent
The architecture is built on a "Surgical Logic" approach—precise, deterministic, and efficient:
The Brain (Agent): Processes natural language intent (e.g., "Find the Pirates of the Caribbean series").
The Nerve (Skill): My custom LazyCat Movie Search Skill queries resource sites and filters metadata.
The Muscle (Downloader): Interfaces with the VueTorrent API on the NAS to execute the task.
Why It’s "Smooth as Silk":
Smart Recognition: It doesn't just find one file; it can identify entire series or specific sequels based on context.
Quality Control: Supports filtering by resolution—720P, 1080P, or 4K.
Traffic Management: I implemented a logic to automatically set the Seeding Ratio to 1.0. This ensures you contribute back to the swarm without choking your NAS upload bandwidth indefinitely.
The "Popcorn" Experience: Once the file hits the NAS, tools like NetEase Popcorn automatically scrape the metadata, creating a beautiful poster wall for the family to watch on any device.
Quick Setup
If you’re running a NAS (optimized for Lazy Cat, but adaptable for others), here’s the drill:
Prepare your Agent: Ensure your assistant (like Hermes) supports external Skill loading.
Install the Skill: Load the "LazyCat Movie Search" Skill.
Link the Downloader: Input your VueTorrent WebUI address and credentials.
Talk is cheap. Show me the code.
I’ve fully open-sourced this Skill. The logic is clean, and you can easily adapt it to other BT clients by tweaking the API calls.
🔗 GitHub Repository: https://github.com/whoamihappyhacking/lazycat-movie-search-skill
Final Thoughts
Now, when my wife wants to watch something, I don't even have to look up from my keyboard. A single voice command to the Agent handles it all.
To me, AI shouldn't just be a chatbot; it should be the "universal interface" between human intent and hardware execution. This project is a small example of how we can use AI to reclaim our time from repetitive digital chores.
If you’re into AI Agents, private clouds, or just want to chat about NAS optimization, feel free to drop a comment or open a PR on GitHub.
Happy Hacking!
Top comments (0)