Skip to content

About

LocalPilot is a Python application that helps nontechnical users discover, install, run, monitor, update and remove local AI models.

Topics

Resources

Stars

1 star

Watchers

1 watching

Forks

Repository files navigation

LocalPilot Banner



LocalPilot Logo

LocalPilot

Run AI on your own computer.

English • العربية (Arabic)

PyPI Version Python Versions License: GPL v3 DOI Platform Documentation

LocalPilot checks your computer, helps you choose a suitable AI model, and guides you through setup.

Website • Beginner Guide • Installation • Model Catalog • FAQ • التوثيق بالعربية • PyPI


💡 What Can LocalPilot Help You Do?

LocalPilot lets you download and run artificial intelligence models directly on your laptop or desktop computer.

You can use it to:

  • Draft and edit writing: Write emails, outline essays, polish documents, and rephrase text.
  • Ask everyday questions: Summarize articles, explain concepts in simple terms, and brainstorm ideas.
  • Get programming help: Explain unfamiliar code, write helper functions, and troubleshoot errors.
  • Keep conversations on your machine: Chat through an offline terminal window where your words are processed locally on your computer rather than sent to cloud servers.

👥 Who Is It For?

  • Everyday users & beginners: You do not need an engineering degree or prior AI experience. LocalPilot handles memory calculations and configuration for you.
  • Privacy-conscious writers & researchers: Keep your personal notes, drafts, and sensitive documents on your own computer.
  • Developers & students: Experiment with modern open-weights models (like Llama, Qwen, Gemma, and Mistral) with reliable local API access.

💻 What Do You Need?

  • Operating System: Linux (Ubuntu, Debian, Fedora, Arch, etc.) or macOS (Apple Silicon M1–M4 or Intel). Windows is supported through Windows Subsystem for Linux (WSL2).
  • Terminal: A standard command-line terminal window where you can paste commands.
  • Internet Connection: Needed to download LocalPilot, download your chosen model, and update model lists. (Once a model is downloaded, chat runs locally without internet).
  • Memory & Storage:
    • 8 GB of RAM and 10 GB of free disk space for lightweight models (1.5B to 3B parameters).
    • 16 GB of RAM or more for standard models (7B to 8B parameters).
  • Python: Python 3.11 or higher is required and is checked automatically during installation.

🚀 Get Started

Recommended Installation (Automated Installer)

Open your terminal window and paste this single command:

curl -fsSL https://raw.githubusercontent.com/fakhruldeen/LocalPilot/main/install-localpilot.sh | bash

What this command does:

  1. Checks that your operating system and Python version are compatible.
  2. Installs LocalPilot into an isolated application folder (~/.local/share/localpilot on Linux, ~/Library/Application Support/LocalPilot on macOS).
  3. Leaves your existing system files untouched and never asks for administrator (sudo or root) passwords.

Alternative Installation Methods (For Python Developers)

If you manage Python command-line tools with pipx, pip, or uv:

# Using pipx (recommended for Python CLI users)
pipx install localpilot

# Using pip in your active environment
pip install localpilot

# Using uv
uv tool install localpilot

👉 For custom scopes, offline bundles, and uninstallation instructions, see the Installation Guide.


🖥️ What Happens After Launch?

Start LocalPilot by typing:

localpilot

This opens a friendly, full-screen dashboard inside your terminal. If this is your first time, the guided setup wizard walks you through:

  1. Computer Capability Check: LocalPilot measures your processor, memory, and graphics hardware, reserving a safe buffer so your operating system and web browser stay smooth.
  2. Choose What You Want Help With: Pick what you want your model to do best—Everyday questions and writing, Coding assistance, or Browse all compatible models.
  3. Recommended Models: LocalPilot shows models that are tested to fit within your computer's safe memory budget.
  4. Review & Approve: You see the exact model name, download size in gigabytes, and storage location. Nothing downloads until you select Set Up Model.
  5. Setup & Verification: LocalPilot downloads the weights file, prepares the supporting software, and runs a quick test question to confirm that everything works properly.
  6. Start Chatting: Select Start Chatting to begin your first conversation immediately.

💬 Everyday Actions

Once set up, you can manage your local AI using simple terminal commands:

Action Command What it does
Open Visual Dashboard localpilot Launches the full-screen interactive terminal dashboard with dedicated tabs for models, tools, search, benchmarks, and health.
Start Chatting localpilot chat Opens an interactive conversation with your installed model. Type /exit to leave.
Run Offline AI Tools localpilot tools <task> Run 18 focused task tools: summarize, rewrite, translate, explain-code, extract, classify, review-table, describe-image, read-image, and more.
Search Local Documents localpilot search ask "q" Offline RAG: Index local folders and get grounded answers with precise file and line citations.
Voice & Speech Chat localpilot voice chat Conversational voice turn loop with offline speech-to-text transcription and text-to-speech reading.
Measure & Benchmark localpilot bench run Test streaming performance (TTFT, tokens/sec) and compare models side-by-side with bench compare.
Health & Self-Repair localpilot doctor --repair Inspect system health, detect misconfigurations, and apply safe, reviewed repairs.
Connect External Apps localpilot connect Interactive wizard to configure VS Code (Continue), OpenCode, Cursor, and custom tools. Manage with localpilot connections.
Stop Model (Free RAM) localpilot stop Unloads the model from memory so all RAM and graphics memory return to other apps.
View Installed Models localpilot installed Lists all models currently downloaded on your computer.
Find Compatible Models localpilot discover Scans the catalog for models that fit your memory budget.
Remove a Model localpilot remove <name> Deletes model weight files from your hard drive to free disk space.
Star a Favorite localpilot favorite add <alias> Adds a model to your favorites list for quick access.

Tip

Leaving Chat vs. Stopping a Model:

  • Typing /exit leaves the chat screen but keeps the model loaded in memory so you can return to it instantly.
  • Running localpilot stop completely unloads the model from memory, freeing up your RAM for games, browsers, or heavy work.

🔒 Privacy, Internet Access & Limitations

  • Local Inference: Supported conversations are processed entirely on your computer through a local loopback address (127.0.0.1). Your messages and files are not uploaded to cloud servers.
  • Internet Requirements: An internet connection is used during installation, when downloading a model for the first time, and when updating the model catalog. Once a model is saved locally, chat runs without an active internet connection.
  • Accuracy Notice: Like all artificial intelligence systems, local models can make mistakes, produce inaccurate answers, or provide flawed code. Always double-check important facts and test code before using it.
  • Hardware Boundaries: Local models share your computer's memory. Running models that exceed your available RAM can make your desktop sluggish. LocalPilot checks your memory first to recommend models that run smoothly.

🔌 Optional: Connect Other Applications

LocalPilot provides a standard, local OpenAI-compatible API endpoint (http://127.0.0.1:8080/v1).

You can connect popular coding assistants and tools:

  • VS Code (Continue extension): Connect your local model to chat with your codebase.
  • OpenCode: Automated configuration generator for local coding tasks.
  • Cursor & Custom Tools: Point any OpenAI-compatible client to your local endpoint.

Run localpilot connect to view guided instructions for each application.
(Note: Connected third-party applications have their own privacy settings and terms).


🛠️ Help & Troubleshooting

  • Download was interrupted: Run localpilot again. The installer checks what was already downloaded and resumes safely.
  • Computer is running slowly: Run localpilot stop to unload the model from memory, or choose a smaller model with localpilot discover.
  • Command not found (localpilot: command not found): Make sure ~/.local/bin is in your terminal's search path. Add export PATH="$PATH:$HOME/.local/bin" to your ~/.bashrc or ~/.zshrc.
  • Access key requested: Most models are open and need no accounts. A few models (such as official Llama or Gemma weights on Hugging Face) require a free read-only personal access token to confirm the publisher's terms.
  • Need diagnostic help: Run localpilot support to generate a safe, redacted diagnostic report you can review and share when asking for help.

📖 Documentation & Guides

Detailed guides and technical references are available across the repository:


👤 Author & Maintainer

Mohamed Fouad Fakhruldeen
Creator & Maintainer of LocalPilot. Working across PMO governance, AI systems, and agile enterprise operating systems.


⚠️ Disclaimer

Warning

Personal Use & Testing: LocalPilot is designed for personal evaluation, learning, and hobbyist exploration. Running local artificial intelligence models uses significant computer memory and processor resources.

This software is provided strictly on an "AS-IS" basis. It is not certified for safety-critical systems, commercial infrastructure, or emergency services. The author assumes no liability for system instability, data loss, or issues resulting from running local models. Always back up important files and review model responses.


📚 Citation

If you use LocalPilot in your research, academic study, or open-source projects, please cite it as:

@software{fakhruldeen_2026_localpilot,
  author       = {Mohamed (Fouad) Fakhruldeen},
  title        = {fakhruldeen/LocalPilot: v0.2.5},
  month        = oct,
  year         = 2026,
  publisher    = {Zenodo},
  version      = {v0.2.5},
  doi          = {10.5281/zenodo.23424540},
  url          = {https://doi.org/10.5281/zenodo.23424540}
}

📄 License

This project is licensed under the terms of the GNU General Public License v3 (GPLv3). See the LICENSE file for details.

About

LocalPilot is a Python application that helps nontechnical users discover, install, run, monitor, update and remove local AI models.

Topics

Resources

Stars

1 star

Watchers

1 watching

Forks

Releases

Contributors

Languages