Getting Started & Cost
Is LocalPilot really 100% free? Are there any hidden fees or limits? +
Yes, LocalPilot is completely free and open-source under the GNU General Public License (GPLv3).
There are zero monthly subscription charges, zero pay-per-token API fees, no premium tier, and no credit card required. Because the AI runs entirely on your own computer hardware, there are no cloud server bills to pass on to you.
Do I need to know how to code or be a tech expert? +
Not at all. LocalPilot was specifically built for students, writers, researchers, and everyday professionals who don't want to fiddle with command-line switches or server setups.
You only need to paste one install command into your terminal once. From then on, typing localpilot opens an interactive visual menu with numbers you can choose from, or a friendly chat screen.
What operating systems are supported? +
LocalPilot currently provides tested and certified support for:
- Apple macOS: macOS Sonoma, Ventura, Monterey on Apple Silicon (M1, M2, M3, M4) as well as modern Intel Macs.
- Linux: Ubuntu 20.04+, Debian 11+, Fedora 38+, Arch Linux, and all standard distributions on x86_64 or aarch64.
How is LocalPilot different from downloading Ollama or LM Studio directly? +
Tools like Ollama and LM Studio are fantastic raw engines, but they expect you to know what models fit, how many gigabytes you have available, and how to manage services yourself. If you pick a model that's too big, your computer can lock up.
LocalPilot acts as your protective co-pilot: it measures your computer's RAM, recommends only models that safely fit, verifies your hardware compatibility, tests inference before saying a model is ready, and allows you to chat immediately without managing servers.
Hardware, Performance & Battery
Will running an AI model freeze or slow down my computer? +
No, because LocalPilot enforces strict RAM safety margins.
Before launching any model, LocalPilot computes your real-time available memory and ensures at least 2.0 to 4.0 GB of RAM remains free for your web browser, video calls, and operating system. If memory is tight, it refuses to load the model and warns you safely.
Do I need an expensive dedicated gaming GPU? +
No. Compact 3B models (like Llama 3.2 3B or Phi-3 Mini) run smoothly on normal Intel or AMD laptop CPUs. If you have an Apple Silicon Mac (M1, M2, M3, M4), its Unified Memory acts as a GPU natively, delivering rapid speeds out of the box.
Does running local AI drain my laptop's battery? +
While the model is actively thinking and streaming words (usually 3 to 10 seconds per response), it engages your processor cores. As soon as it finishes printing the answer, it returns to an idle sleep state using almost zero battery.
On modern Apple Silicon MacBooks, battery impact is so small that you can easily chat for several hours on a single battery charge.
Will my laptop get hot or make loud fan noise? +
For quick questions and email drafts, generation takes just a few seconds, so your laptop fans will rarely spin up. If you ask the model to analyze a 20-page document or write an entire novel chapter on an Intel laptop, your fans may briefly turn on while processing, which is completely normal.
Privacy, Security & Data Safety
Can anyone see my chats, prompts, or documents? +
Nobody. Not even LocalPilot.
When you chat with a local model, your text never leaves your computer's RAM. There are no servers, no telemetry, no tracking analytics, and no logs uploaded anywhere. You can verify this yourself by turning off your Wi-Fi completely and chatting in Airplane mode!
Are downloaded model files safe from viruses or malware? +
Yes. AI model files (.gguf) are static mathematical data files containing weight matrices (millions of numbers), not executable programs like .exe or .sh scripts. They cannot execute unauthorized instructions or install software on your machine.
Additionally, LocalPilot verifies SHA-256 cryptographic hashes on official downloads to guarantee that the file matches the exact bytes published by verified research labs (Meta, Google, Microsoft).
Can LocalPilot touch, delete, or read my personal files? +
No. LocalPilot operates strictly within its own designated application folder (~/.local/share/localpilot on Linux or ~/Library/Application Support/localpilot on macOS).
It runs under your normal user account without administrator or root privileges, meaning it physically cannot modify system settings or touch files outside its managed directory.
Models, Downloads & Storage
How large are the model files, and how long does downloading take? +
Compact models like Llama 3.2 3B or Qwen 2.5 3B are around 2.0 GB to 2.3 GB in download size. On a typical 100 Mbps home Wi-Fi connection, download takes about 2 to 4 minutes.
Larger 8B models (like Llama 3.1 8B) are roughly 4.7 GB, taking around 5 to 8 minutes to download once. Remember: you only download a model once, and then it is yours forever.
Which model should I download first as a complete beginner? +
We recommend starting with Llama 3.2 (3B):
It is lightweight, extremely responsive, and excels at everyday tasks like summarizing notes, proofreading emails, and answering general questions.
How do I delete a model and reclaim hard drive space? +
You have total control over your disk space. To view stored models and their file sizes, run:
To completely delete a model file and return its gigabytes to your hard drive, simply run:
How do I update models when newer revisions are released? +
Running localpilot discover automatically queries online registries and checks for new compatible candidates and updated revisions.
If an updated model is available through Ollama or Hugging Face, LocalPilot downloads the new version to a staged recovery folder first, tests it, and ensures your previous working version remains intact if anything goes wrong.
Runtimes, Ollama & Docker Simplified
What is an "inference runtime" or "backend"? +
Think of a model file like a digital cookbook. The "inference runtime" (like llama.cpp) is the chef that reads the recipes, mixes the numbers, and serves the finished text. LocalPilot automatically chooses and configures the best chef for your specific computer.
What is Docker, and do I need to install it? +
No, you do not need Docker at all.
Docker is a tool that software developers use to isolate apps in virtual containers. LocalPilot runs 100% natively on macOS and Linux without containers. If you don't have Docker installed, you can completely ignore it.
What if I already have Ollama or LM Studio running? +
LocalPilot follows a strict Golden Rule of Ownership: it will never stop, restart, or delete files belonging to external software.
If it discovers a running Ollama service, it can connect to it to run models, but it will never interfere with your existing models or settings.
Troubleshooting & Support
How do I check if my computer is healthy and ready? +
Run the built-in diagnostic doctor:
It performs a non-invasive scan of your RAM, processor, storage, and acceleration libraries (Metal / CUDA) and tells you if anything needs attention.
How do I completely remove LocalPilot if I change my mind? +
LocalPilot provides a clean 1-command self-uninstall:
This cleanly deletes the application virtual environment, desktop shortcuts, and stored models, leaving zero leftover files or background services behind.
How do I get help or submit a diagnostic report safely? +
If you run into an unexpected issue, generate an anonymized, redacted support bundle:
It exports hardware diagnostics and error logs with all personal paths and names stripped out so you can safely paste it into a GitHub Issue for developer assistance.