Rippletide Context Graph

Rippletide Context Graph

The Context Graph is a persistent context graph that gives your AI agents memory across conversations.

When to use it

• Your agent needs to recall past interactions or user preferences
• RAG doesn't find the right data, even after many iterations it is stil plateuing
• Your agent hallucinates since it doesn't have the elements to answer
• You don't want each new session to start from zero, you need previous decisions made and  context established
• You want to plug memory into any MCP-compatible client (Cursor, Claude, VS Code)

What the Context Graph does

The Context Graph is a structured, persistent knowledge store that sits behind your Azure AI agent. It stores facts, decisions, preferences, and entity relationships as a queryable graph — far more powerful than simple RAG or flat conversation history. Your agent automatically calls memory tools during conversations (remember, recall, relate, invalidate) with no manual wiring required.

Key features

• Native MCP integration: plug into GitHub Copilot, VS Code AI extensions, and Azure-hosted agents
• 7 MCP tools: remember · recall · get_context · list_entities · relate · invalidate · switch_agent
• Structured entity types that adapt to your use case: Person, Organization, Product, Decision, Action, Concept, Location...
• Typed relationships that adapt to your use case: works_at, approved, declined, resulted_in, caused, and more
• Isolated graphs per agent: memories from one Azure workload never leak into another
• Confidence scoring: hard facts (1.0) vs. inferred context (0.5–0.8)
• Full explainability: invalidate stale memories without deleting history - full traceability
• Versioned data modifications to keep track

Who it's for

• Microsoft ISVs and Azure-native SaaS teams building AI agents
• Enterprise developers extending GitHub Copilot or Microsoft 365 Copilot with long-term memory
• Azure AI Studio teams adding structured context to their agent pipelines

How to get started

1. Sign up at app.rippletide.com
2. Add the Rippletide MCP server to your Azure agent or VS Code config (one JSON block)
3. Paste the system prompt into your agent
4. Your agent starts remembering - automatically, across every session

Works with: GitHub Copilot · VS Code · Azure AI Studio · Azure OpenAI Service · any MCP-compatible client