Up and running in five minutes
From zero to a context-aware AI agent: create a project, add topics, connect your tool. No infrastructure to configure.
How it works
Six steps from account creation to a fully context-aware agent.
Create an account
Sign up for free, no credit card required. The free plan gives you one project and up to ten topics, enough to see the full flow in action.
Create a project
A project is the container for your knowledge base. Name it after the product, client, or context you are working on. Each project gets its own API key and isolated MCP server.
Add your first topics
Write down what your agent should know: architectural decisions, operational instructions, domain context, key concepts. Each topic can be toggled on or off for the MCP endpoint independently.
Copy the MCP endpoint
From the project dashboard, copy the endpoint URL and your API key. You will use these to configure your AI tool.
Configure your AI tool
Paste the endpoint into your client config: Claude Code, GitHub Copilot, Cursor, or any other MCP-compatible tool. One-time setup; the agent picks up all changes automatically from that point on.
Your agent has context
From now on, your agent queries the knowledge base before working. No more re-explaining decisions, context, or constraints at the start of every session.
Next steps
Once the agent is connected, explore what else meridiaan can do.
Ready to give your agent a memory?
Free plan available, no credit card required.