Eidolon Chat

Eidolon Chat is the core of the Eidolon Home AI ecosystem. A complete AI assistant that runs entirely on your hardware.

Summary

• a complete local AI system
• modular and scalable
• designed for real-world use

It combines:

• independence
• performance
• flexibility

Eidolon Chat is a local-first artificial intelligence system designed to run entirely on the user’s device, providing a powerful, modular, and privacy-focused conversational environment.

Unlike traditional cloud-based solutions, Eidolon does not require continuous data transmission to external servers:
all interactions are processed locally, ensuring full control, data ownership, and operational independence.

Local Architecture and Privacy

Eidolon operates under a local-first architecture:

• No data is transmitted externally by default
• Conversations remain on the user’s device
• Processing is handled by locally executed LLMs
• External connections are optional and fully user-controlled

For advanced tasks, users can optionally connect Eidolon to cloud-based models via API, but this is:

• explicitly enabled
• fully configurable
• never automatic

AI Engine and Scalability

Eidolon Chat is designed to adapt to the available hardware.

Currently supported models include:

• Gemma 3 4B → lightweight, fast, suitable for CPU or low-end GPU
• Gemma 3 12B → more powerful, improved depth and response quality

The system is modular and continuously evolving:
additional models will be integrated over time, allowing users to choose the best balance between performance and capability.

Core Features

🧠 Full AI Chat

A complete conversational system with:

• Context-aware responses
• Multi-turn interaction
• Natural, continuous dialogue

🎭 Four Built-in Personalities

🎭 Thomas — The Sarcastic Analyst Razor-sharp wit and dry humor. Ideal for technical analysis and code review.

🌟 Natasha — The Visionary Idealistic and creative. Perfect for brainstorming and unconventional ideas.

⚖️ Jason — The Balanced Mind In medio stat virtus. Measured reasoning and balanced judgment for complex decisions.

💙 Sue — The Empath Warm and emotionally intelligent. Ideal for personal support and coaching conversations.

🧩 Persistent Memory

The system includes local persistent memory:

• Storage of relevant interactions
• Long-term contextual awareness
• Behavioral continuity over time

Memory is structured to ensure coherence and consistency across sessions.

👁️ Computer Vision

Eidolon can process and interpret images:

• Automatic image description
• Contextual visual understanding
• Integration of visual input into conversation

📄 Document Analysis

Support for document processing:

• Information extraction
• Summarization and restructuring
• Context-aware interpretation

🧠 PIF (Prompt Instruction Framework)

An advanced system for guiding model behavior:

• Role and objective definition
• Fine-grained response control
• Advanced interaction customization

🌐 Internet Search (Optional)

Controlled access to online information:

• Manually activated
• Used for updates and fact-checking
• Integrated into the conversational flow

🎤 Voice Interaction

Voice-enabled communication:

• Speech-to-text input
• Text-to-speech output
• Natural, hands-free interaction

Optional Cloud Integration

For complex or high-demand tasks:

• API-based connection to cloud LLMs
• Fully manual activation
• No automatic data transmission

This allows users to:

• operate locally for everyday use
• scale to more powerful models only when needed

• 🔗 Moltbook ❌ (available in Social Edition)
• ☁️ Cloud API bridge — optional connection to OpenAI, Anthropic, Google and more

System Philosophy

Eidolon Chat is built around a core principle:

the user remains in full control of the AI system

This is reflected in:

• local execution
• modular design
• transparency
• strong privacy guarantees

offering a practical alternative to centralized, cloud-based AI platforms.

System requirements: Windows 10/11, 16 GB RAM (32 GB+ recommended), GPU with 6+ GB VRAM, 15 GB free disk space.