How does the Recall skill handle large amounts of data?

It uses context-aware result limits that automatically reduce detail and result counts as your context window fills up, preventing token overflow.

Do I need a cloud account to use Recall?

No, the primary graph search works locally out-of-the-box. Cloud semantic search via mem0 is an optional feature requiring an API key.

Can I search across different projects?

Yes, using the --global flag allows you to retrieve best practices and patterns that have been shared across all your projects.

What is the benefit of a graph-first search?

Graph search allows for explicit relationship queries (like 'find what technology X uses') which is often more precise than keyword search for architecture.

Knowledge Graph Recall

Name: Knowledge Graph Recall
Author: yonatangross

byyonatangross

•

生产力与工作流

Searches and retrieves architectural decisions, coding patterns, and project context from a structured knowledge graph.

关于

Knowledge Graph Recall is a sophisticated context-retrieval skill designed for Claude Code that enables developers to access past technical decisions and established patterns instantly. By leveraging a graph-first architecture, it provides zero-configuration local search for entities and relationships, with optional cloud-based semantic enhancement via mem0. The skill intelligently manages context window pressure by automatically summarizing results based on token usage, ensuring Claude always maintains the most relevant project background for complex development tasks.

主要功能

27 GitHub stars
Context-Aware Scaling: Automatically adjusts result density based on current context window usage.
Scoped Searching: Filter results by specific agent IDs, categories, or global cross-project best practices.
Relationship Mapping: Identifies and displays links between technologies, patterns, and architectural decisions.
Graph-First Architecture: Local, zero-config entity and relationship search using MCP memory.
Semantic Cloud Enhancement: Optional mem0 integration for deep semantic similarity search.

使用场景

Retrieving past Architectural Decision Records (ADRs) to ensure technical consistency.
Recalling project-specific coding patterns and anti-patterns during refactoring.
Querying cross-project best practices to implement standardized solutions like pagination or auth.

关于

主要功能

27 GitHub stars
Context-Aware Scaling: Automatically adjusts result density based on current context window usage.
Scoped Searching: Filter results by specific agent IDs, categories, or global cross-project best practices.
Relationship Mapping: Identifies and displays links between technologies, patterns, and architectural decisions.
Graph-First Architecture: Local, zero-config entity and relationship search using MCP memory.
Semantic Cloud Enhancement: Optional mem0 integration for deep semantic similarity search.

使用场景

Retrieving past Architectural Decision Records (ADRs) to ensure technical consistency.
Recalling project-specific coding patterns and anti-patterns during refactoring.
Querying cross-project best practices to implement standardized solutions like pagination or auth.