ContinumAI: The Memory That Survives When the Model Changes

Description and use

ContinumAI preserves context across AI model switches, enabling continuous workflows without losing conversation history.

Problem

Switching AI models after extensive work causes loss of context, preferences, and unfinished tasks, making each new model feel like starting over.

Outcome

ContinumAI demonstrates that persistent memory can survive model changes, enabling continuous AI workflows across multiple providers.

Workflow

• 1. Integrate multiple model providers via adapters.
• 2. Store conversation exchanges as nodes in a PostgreSQL database with prev_id links and checkpoints every 5 turns.
• 3. Orchestrate requests through a Node.js/Express gateway that selects the appropriate provider.
• 4. Visualize the memory graph in a React/TypeScript frontend.
• 5. Allow users to inspect, manage, and transfer relevant context across models.

Evidence

ContinumAI is a multi-model memory and gateway project designed to help AI conversations maintain continuity across different models.

Source

The idea behind the project is simple: The model can change. The memory shouldn't have to start from zero.

Source

The prototype showcases the ContinumAI memory graph, where conversation nodes, model providers, and relationships can be explored visually.

Source

The project explores a modular architecture built around: Multi-provider integration, Persistent memory, Gateway orchestration, Memory visualization, Extensible architecture.

Source