Agents Machine

Memory Browser

Explore and manage your AI's knowledge base visually.

The Memory Browser provides a visual interface to explore, search, and manage all stored memories in the Qdrant vector database.

🔍 Semantic Search

Natural language queries find related content via vector similarity.

📁 Category Filter

Filter by architecture, coding-rules, anti-patterns, service-context.

🏷️ Tag Filter

Search by specific tags for precise results.

Categories & Types

CategoryPurposeExamples
architectureSystem design decisions"Use PostgreSQL for relational data"
coding-rulesCode style and conventions"Always use TypeScript strict mode"
anti-patternsThings to avoid"Never use any type"
service-contextProject-specific knowledge"Auth service uses JWT with refresh tokens"
TypePurpose
ruleEnforced coding or architecture rule
memoryGeneral project knowledge
anti-patternKnown bad practice to avoid
guideHow-to or reference
decisionRecorded design decision

Managing Memories

Store memories via the desktop chat, IDE, or MCP tools:

Store in memory: Our API uses rate limiting of 100 req/min per user.
Tag it as: api, rate-limiting, infrastructure
Category: service-context

Click any memory to view its contents, category, tags, and metadata.

Select memories and delete them. Use delete_memory with IDs for batch deletion via MCP.

On this page