Ainara /aɪˈnɑːrə/ (n.) [Basque origin]: 1. A feminine given name meaning "swallow" (the bird) or "beloved one". [..] Associated with spring, and the beginning of life.
Ainara is a modular AI integration platform that reimagines human-computer interaction through natural conversation, made with components which work together to create intelligent companions that collaborate helping with tasks, generating insights, and transforming how people work with technology through voice and intuitive interfaces.
It differentiates itself from other projects with its "user-first" philosophy. Eg. AI skills/tools can be both locally in the user's system, using the Orakle server approach developed by this project, or be accessed remotelly via the MCP protocol. The project also emphasizes a "local-first" AI approach via Ollama full integration, although this is not strictly enforced, allowing users to select from over 100 LLM providers via the excellent LiteLLM library.
Finally, this project creates a truly AI collaborating-companion experience. Conversations are not session-based; user interactions with the LLM are recorded permanently as a continuous conversation (though users can choose to disable the memory feature at any moment).
All interaction data remains private on the user's system.
UPDATE July 24th, 2025: This is the 13th video in my series featuring Ainara. This time is featuring a cloud LLM model, Grok 3 mini provided by xAI. The video shows the full process of installing, configuring Ainara and then a short demo showing the capabilities of Ainara detecting the user intention and applying the corresponding skills (or tools). MCP is now integrated in Ainara that allows to integrate external services in the AI like Google Maps, like is featured in this video.
A REST API server that provides:
- Extensible skills system
- Working client-side, all the skills and a live conversation can be hot-swapped between LLM providers
- MCP compatible for third party servers.
A modern desktop-integrated application that provides:
- Native integration with system features
- Intuitive, minimalistic AI interaction interface
- Rich graphical interface for chat interactions
- Real-time skill execution feedback
- System tray presence for quick access
- Cross-platform support (Linux, Windows, macOS)
CLI chat interface for Ainara. Right now is outdated and needs some further work.
List of the currently available skills already integrated in the Ainara AI Assistant Framework:
- Finance Stocks: Get stock market information.
- Search Engines (Google, Metaphor, NewsAPI, Perplexity, Tavily): Perform combinated web searches using various search engines.
- System Clipboard: Read and write the system clipboard.
- System Finder: Intelligent file search with LLM-assisted disambiguation and location reveal.
- Time Weather: Get weather information.
- Tools Calculator: Evaluate non-trivial mathematical expressions.
- Clone the repository:
git clone https://github.com/yourusername/ainara.git
cd ainara- Install dependencies (if you need a python virtual environment we suggest you to create it in the
/venvsubdirectory):
pip install -r requirements.txt# Start the backend servers using the services script
python bin/services.py --health-check
# In another terminal, once the backend services are healthy, start the Polaris frontend in dev mode
npm install # Only needed first time
npm run devThe services.py script manages both the Orakle and PyBridge servers, handling their startup, health monitoring, and shutdown.
You can build the backend servers using PyInstaller:
# From project root
pyinstaller scripts/pyinstaller/servers.specThis creates standalone executables for both Orakle and PyBridge servers.
To build the complete Ainara package for your platform:
# From the polaris directory
npm run build:linux # For Linux
npm run build:win # For Windows
npm run build:mac # For macOSThis will:
- Build the backend servers using PyInstaller
- Package the Electron frontend
- Create a complete distributable package
Polaris is the recommended way to interact with Ainara, providing a modern, desktop-integrated experience.
Polaris features:
- System tray integration for quick access
- Minimalistic, non-intrusive interface
- Typing mode for direct text input
- Real-time skill execution and feedback
- Seamless integration with Orakle skills
Polaris provides a configuration wizard to easily handle the backend/frontend configurations settings.
Ainara stores its configuration in platform-specific locations:
- Windows:
%APPDATA%\ainara\ainara.yaml - macOS:
~/Library/Application Support/ainara/ainara.yaml - Linux:
~/.config/ainara/ainara.yaml
Inside the polaris subdirectory there's a specific polaris.json file with the specific frontend settings.
The variables used by LiteLLM can be still used for API keys and model configuration:
OPENAI_API_KEY: For OpenAI servicesANTHROPIC_API_KEY: For Anthropic services- Other provider-specific variables as documented by LiteLLM
These environment variables are optional and only needed if you want to override settings in the configuration file.
- Python 3.12
- Dependencies listed in requirements.txt
- Orakle API server running locally or on network
- Optional local LLM server (compatible with OpenAI API format)
Dual-licensed under LGPL-3.0 (open source) and commercial terms (contact email)
The Ainara Project now has its own Solana cryptocurrency token, CA: HhQhdSZNp6DvrxkPZLWMgHMGo9cxt9ZRrcAHc88spump
The Ainara Project will use this token in a coming up distributed app store platform, for a new type of applications called Nexus as described in: https://ainara.app/AINARA_NEXUS_APPS_PLATFORM_V1_1.pdf
Everyone's invited to join this project - developers, designers, sponsors, testers, and more! My ultimate goal would be to create an open, community-driven AI companion/assistant that achieves for the emerging open source AI tools what Linux did for Unix: a widely adopted, powerful, and endlessly customizable assistant that empowers users and developers alike.