LocalAI-AngelAngel AI
The assistant interface and starting point for everyday AI workflows.
LocalAI-Angel Architecture Overview
The project connects the Angel assistant experience, model inference, modular tools, persistent storage, and supporting services.
One project, three connected layers
LocalAI
The inference and API layer for supported models and local services.
Angel Nexus
The modular tools and integration environment, with capabilities depending on the current build.
LocalAI-Angel is actively evolving. Supported capabilities and performance depend on the build, configuration, hardware, and service readiness.
Install, start, and stop Angel on Windows
Use the Angel Control Center for the easiest setup and everyday operation. Docker Desktop must be installed and running before you install, start, or manage the platform.
Before you begin: Open Docker Desktop and wait until its engine reports that it is running. Keep Docker running while installing and while using Angel. The first installation can take a while because services and the default model may need to download or build.
Download LocalAI-Angel
Download the ZIP from GitHub and extract the repository to your preferred folder. For example:
D:\LocalAI-Angel
Start Docker Desktop
Install Docker Desktop if you do not already have it. Launch Docker Desktop and wait until the Docker engine is fully running before opening the installer or Control Center.
Open the Control Center
Open the project's dist folder and double-click the Control Center executable:
dist\AngelControlCenter.exe
Keep the project folder structure intact so the Control Center can find the batch scripts.
Install or repair — first-time setup
In the Control Center, click INSTALL / REPAIR. This runs the installation workflow to prepare the LocalAI-Angel services and required components. Follow the on-screen progress and allow downloads and builds to finish. Do not close Docker Desktop during installation.
Start Angel for everyday use
Once installation has completed, open dist\AngelControlCenter.exe and click START ANGEL. The Control Center starts the configured Docker Compose services. Wait for startup to complete, then open the local interfaces below.
Stop Angel when finished
Open the Control Center and click STOP ANGEL. This stops the configured services. Persistent application data and model volumes are intended to remain in place, so you do not need to reinstall every time you want to use Angel again.
Everyday workflow: Start Docker Desktop → open
dist\AngelControlCenter.exe → click START ANGEL. When finished, click STOP ANGEL. Use INSTALL / REPAIR for the initial setup or when you need to repair the installation.First launch: allow internet access for the initial image, source, and model downloads. The default model is
qwen2.5:3b and is not bundled in the ZIP. Download and build time depends on your internet connection and hardware. Never share API keys, tokens, or private .env files.Open Angel, LocalAI, and Nexus
After the services finish starting, open these local interfaces on the computer running the containers. These addresses are not public internet URLs.
Angel AI · Home Base
Assistant chat, history, summaries, knowledge, and related features.
http://127.0.0.1:8765/
LocalAI · Model management
Model gallery and inference API. Add supported models here as needed.
http://127.0.0.1:8080/
Angel Nexus · Modules
Browse and manage the Nexus module and repository workflows.
http://127.0.0.1:8877/
How model selection works
- Angel uses LocalAI when LocalAI has a chat model available.
- If LocalAI has no chat model, Angel can fall back to the provisioned
qwen2.5:3bmodel. - The default Ollama model is stored separately from LocalAI's model gallery; it is not automatically copied into LocalAI.
- Free-form AI reasoning requires a language model. Some deterministic date/time, weather, and knowledge routes can work without one, depending on configuration.
If a local URL does not open, confirm Docker Desktop is running, click START ANGEL in the Control Center, and wait for the services to finish initializing. Check the Control Center output or use the repository's diagnostic utility if needed.
Control Center Video Guides
Use these short tutorials to follow the installation, startup, and shutdown workflow. Each guide is collapsed by default so the animations do not all play at once.
▶️ 1. Install or repair LocalAI-Angel
Start Docker Desktop first, then open dist\AngelControlCenter.exe and choose INSTALL / REPAIR. Keep Docker running until setup and checks finish.
▶️ 2. Start Angel using the Control Center
Make sure Docker Desktop is running, launch dist\AngelControlCenter.exe, and click START ANGEL. Wait for the services to initialize before opening the local interfaces.
▶️ 3. Stop Angel using the Control Center
Open the Control Center and click STOP ANGEL when you are finished. Stopping the services is not intended to remove persistent models or application data.
Explore the project
Source code ↗
Browse the current code, launchers, configuration, and history.
Angel AI Harness ↗
Explore the assistant harness and starter components.
LocalAI documentation ↗
Learn about model serving, supported backends, APIs, and configuration.
Some features may be experimental, hardware-dependent, or still in development. This repository integrates upstream LocalAI; review license files and attribution notices for the project and its individual components.