Meko organizes agent data around a few key concepts. Understanding these will help you get the most out of the platform.

{{< card-group >}}
  {{< card
    href="architecture/core-concepts/datapacks/"
    title="Datapacks"
    desc="The fundamental storage primitive for multi-agent systems. One datapack per application, or one project. It is the isolation unit for agents working together. Each datapack bundles a database, the knoweledge base, agent memories, collective memories, shared knowledge, MCP server, and credentials."
  >}}
  {{< card
    href="architecture/core-concepts/memory/"
    title="Memory"
    desc="Extracted memories from conversations are specific pieces of information or preferences that are explicitly stored persistently for the long term. Memories are stored as entity and relationship, powered by Meko's graph storage."
  >}}
  {{< card
    href="architecture/core-concepts/knowledgebase/"
    title="Knowledge bases"
    desc="Information repository that acts as a single source of truth for agents, grounding decisions, reducing hallucinations, and ensuring consistent responses. Uses semantic search and RAG to understand query intent, not just keywords. Documents are chunked, embedded, and indexed for retrieval."
  >}}
  {{< card
    href="architecture/core-concepts/traces/"
    title="Traces"
    desc="Meko provides end-to-end visibility into every conversation, from initial prompt through LLM chain-of-thought, including memory and knowledge, the associated API and SQL calls, the queries executed, and the results retrieved."
  >}}

{{< /card-group >}}

<!--
## Four pillars of value

### Accelerate agent development

Stop stitching together data across multiple silos. Meko exposes agent-native primitives, such as "add knowledge," that abstract away storage complexity. A PDF can live in object storage, in real-time updates to SQL tables, or in both, without you having to model the implementation. Your code works with agent concepts, not database schemas, dramatically reducing development time.

Meko is a fundamentally multi-agentic, open-source data layer that enables agents to share knowledge and memories, integrates with the AI ecosystem through open APIs, and is available as both a serverless managed service and for private cloud deployment.

### Higher performance, more efficient agents

Meko's AI-tailored serverless architecture and multi-tenant design handle the variable resource utilization inherent to agentic applications. Agents consume minimal resources when idle and burst instantly when active. Meko optimizes common AI operations like parsing, chunking, and entity extraction with cost-effective models, giving you accuracy versus cost tradeoffs that generic databases cannot provide.

### Automatic data tiering lowers storage costs

Conversations are fundamental to agents but expensive to store in real-time databases. Meko keeps all conversation history queryable while transparently tiering older messages to object storage. Configurable retention policies purge data beyond your needs. This applies across all data models, keeping hot data fast and cold data cheap without manual intervention.

### Full chain-of-thought traceability

End-to-end observability shows exactly which data operations your agents performed, the latencies at each stage, and the token costs per interaction. Meko collects traces for the complete chain of thought, giving you the explainability production AI applications require. Debug performance issues and optimize costs with visibility into your data layer's actual behavior.
-->