Meko for Amazon Bedrock AgentCore
AgentCore Memory remembers this agent’s session, but not what another agent on your account resolved last week.
Over MCP, Meko provides AgentCore with:
Agent-scoped memory
What your agent learns persists beyond the session and is readable via every AI tool you connect to.
A shared knowledge base
Documents and promoted conclusions are searchable by all agents and teammates with access to the datapack.
A trace of every read and write
Each conversation is stored with its reasoning, so you can see what an agent did and why.
Every agent on the account gets shared context
AgentCore Memory covers a single agent session. A Meko datapack sits alongside it so every Runtime-hosted agent reads and writes the same memory and knowledge base, whatever framework runs inside.
Integration is code-level MCP from your Runtime entry point, connecting directly or through a Gateway target. Operations covering memory, conversations, knowledge base, datapacks, artifacts, and observability are available for you, and your code calls the same endpoint directly when an operation must happen every time.
How Meko Benefits AgentCore
Start every run from what the fleet knows.
The entry point calls the memory and knowledge base search tools before the agent reasons, so it begins with prior incidents and runbooks to enrich its context.
Log every run with its reasoning.
Conversation tools record each turn with its input, output, reasoning, and plan, readable after the session ends from any MCP client.
Run any framework inside.
Strands, LangGraph, CrewAI, or hand-rolled: AgentCore is a deployment shell, and the Meko calls are the same plain MCP either way.
Connect directly, or route through Gateway.
Call the endpoint with your API key for a single integration, or register Meko as a Gateway target when you already manage several tools that way.
Connecting Meko in AgentCore
- Create a Meko API key and datapack in the Meko portal.
- In your Runtime entrypoint, connect your framework’s MCP client to https://mcp.mekodata.ai/mcp with an Authorization: Bearer <your-api-key> header and a User-Agent header (e.g. meko-agentcore/1.0) – Meko’s edge rejects requests without one.
- Deploy to AgentCore Runtime as normal. The Meko calls are identical to a plain Python script.
- Optional: Register the same endpoint as a Gateway target alongside your other Gateway tools.