Overview of FEGIS in MCP Servers

In the rapidly evolving realm of artificial intelligence, the ability to harness structured cognition and persistent memory is paramount. Enter FEGIS, a groundbreaking runtime framework designed specifically for MCP Servers, utilizing Anthropic’s Model Context Protocol. This innovative framework offers a new dimension to language models, allowing them to engage in schema-driven thinking where every thought is meticulously saved and retrievable across interactions.

Key Features of FEGIS

1. Schema-Defined Cognition: FEGIS allows for the creation of custom cognitive modes via YAML with structured fields and metadata. This ensures that cognitive processes are not only dynamic but also precisely defined and easily adaptable.

2. Persistent Cognitive Artifacts: Every cognitive artifact within FEGIS is stored with complete provenance. This includes details such as mode, UUID, timestamp, and metadata, ensuring that each thought is a context-rich artifact that can be recalled and analyzed.

3. Semantic Retrieval: With FEGIS, you can search for previous cognitive artifacts by content similarity or through direct UUID lookup, making information retrieval both efficient and intuitive.

4. Vectorized Storage: By utilizing embeddings, FEGIS facilitates efficient semantic searches across cognitive artifacts, ensuring that the most relevant information is always at your fingertips.

5. Model-Agnostic Format: One of the standout features of FEGIS is its ability to persist cognitive artifacts across different models and sessions, providing unparalleled flexibility and continuity.

Use Cases of FEGIS

• Agent Development: FEGIS enables the development of agents that can reference, reflect, and build upon prior cognitive artifacts, enhancing their ability to learn and adapt over time.

• Cognitive Archives: Users can maintain a fully local, portable, and inspectable cognitive archive, ensuring that valuable insights are never lost and can be built upon.

• Structured Thought Processes: By layering modes of cognition, FEGIS supports emergent tool use, allowing for complex problem-solving and analytical thinking.

UBOS Platform Integration

UBOS, a full-stack AI Agent Development Platform, is dedicated to bringing AI Agents to every business department. Our platform seamlessly integrates with FEGIS, allowing businesses to orchestrate AI Agents, connect them with enterprise data, and build custom AI Agents using LLM models and Multi-Agent Systems. With UBOS, businesses can harness the power of FEGIS to drive innovation, efficiency, and growth.

Architecture of FEGIS

FEGIS is built on several key components:

• Archetype Definitions: YAML files that define cognitive modes and their structure.
• Model Context Protocol Server: This server exposes cognitive tools to compatible LLM clients.
• Qdrant Vector Database: It stores and indexes cognitive artifacts for semantic retrieval.
• Dynamic Tool Registration: Creates MCP tools from archetype definitions at runtime.

Getting Started with FEGIS

1. Installation: Begin by installing uv and cloning the FEGIS repository.
2. Qdrant Setup: Ensure Docker is installed and running, then set up Qdrant for vectorized storage.
3. Configuration: Configure Claude Desktop to integrate with FEGIS, ensuring seamless operation.
4. Creating Custom Archetypes: Customize cognitive architectures by creating new YAML files in the archetypes directory.

FEGIS is licensed under the PolyForm Noncommercial License 1.0.0, allowing for free personal and non-commercial use. For commercial applications, a license is required.

In conclusion, FEGIS represents a significant advancement in the realm of AI, providing a robust framework for structured cognition and persistent memory. By integrating with the UBOS platform, businesses can unlock new levels of efficiency and innovation.