OpenViSTAI · Agents · Blockchain · Crypto
Menu
Local & Self-hosted AI

Dalai

Dalai is a self-hosted application for installing and running LLaMA and Alpaca language models on a personal computer. It includes a web interface, uses llama.cpp, llama-dl CDN and alpaca.cpp, and documents JavaScript and Socket.io APIs.

Independent OpenViST project profile, based on linked sources.

Dalai — OpenViST project profile in Local & Self-hosted AI.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Dalai when you want to install and run a supported model locally and can provide the required software, memory and storage.

Who it is for

  • People who want to experiment with LLaMA or Alpaca models on a personal computer.
  • Developers interested in a hackable local web app or its documented JavaScript and Socket.io APIs.

Documented uses

  • Install a supported LLaMA or Alpaca model.
  • Launch a local web interface for using an installed model.
  • Choose a model size with documented memory and disk requirements in view.
Before you choose

Requirements & limits

  • The documented macOS setup requires Node.js 18 or later; the Docker Compose route requires Docker installed and running.
  • The documentation lists Linux, Mac and Windows support, alongside a Docker Compose setup.
  • Estimated memory ranges from about 4 GB for 7B models to 32 GB for 65B models.
  • Quantized LLaMA storage ranges from 4.21 GB for 7B to 40.88 GB for 65B; retaining the original files increases total storage needs.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Install a model and launch Dalai's local web interface on macOS

Install a LLaMA or Alpaca model and start Dalai's web UI on a Mac.

Before you start

  • macOS computer
  • Node.js 18 or later
  • Sufficient memory and disk space for the selected model

Steps

  1. Install Node.js 18 or later if it is not already installed.

  2. Install a model with the command below or the command below; the documented model size can be changed.

    Documented command
    npx dalai llama install 7B
    Documented command
    npx dalai alpaca install 7B
  3. Start the web UI with the command below.

    Documented command
    npx dalai serve
  4. Open http://localhost:3000 in a browser.

Expected outcome

Dalai's local web interface is available at localhost:3000 for the installed model.

Guide limitations

  • The documentation estimates about 4 GB of memory for a 7B model and lists additional disk requirements for model files.
  • If the install commands exit immediately, the documentation suggests installing Homebrew dependencies cmake and pkg-config, updating npm, then retrying.
Source-based practical guide

Install an Alpaca model with Docker Compose

Build Dalai with Docker Compose, install an Alpaca model and start the service.

Before you start

  • Docker installed and running
  • Sufficient disk space and memory for the selected model

Steps

  1. Build the service with the command below.

    Documented command
    docker compose build
  2. Install the model with the command below.

    Documented command
    docker compose run dalai npx dalai alpaca install 7B
  3. Start the service with the command below.

    Documented command
    docker compose up -d
  4. Open http://127.0.0.1:3000.

Expected outcome

The documented Docker Compose workflow stores models in the `./models` folder and serves the site at 127.0.0.1:3000.

Guide limitations

  • The guide requires Docker to be installed and running.
  • Its example installs Alpaca 7B; model memory and storage needs depend on the selected size.
Evidence and transparency

Sources & review

Source record details
Source availability
Active
Sources checked
15 October 2026
Original website title
cocktailpeanut/dalai · Official homepage title · retained in its source language