meridiaanProject memory for the age of AI agents
meridiaan is a knowledge base infrastructure built for agentic workflows, ready to use out of the box, exposed as an MCP server.
The problem is familiar to anyone who has worked with AI coding agents: the agent is capable, but it doesn't know your project. Every session starts from zero. You re-explain the architecture, the conventions, the decisions made last week. The agent is smart: it just has no memory.
The obvious fix is an MCP server that exposes your project context. But building and hosting one takes time, infrastructure, and maintenance. For most developers, that overhead isn't worth it.
meridiaan was built to remove that friction. Create a project, add topics (architectural decisions, operational notes, domain concepts), and connect any MCP-compatible client. Your agent gets context immediately. No server to deploy, no pipeline to manage.
The knowledge base grows with the project. It's not documentation you write once and forget: it's a living source of truth that you update as things change, and that every member of your team and every agent can query in real time.
What we believe
Built for developers
meridiaan was created by a developer, for developers. The language, the workflow, the defaults: everything is shaped around how engineers actually think and work.
No overhead
We believe the best tool is the one that gets out of your way. No boilerplate, no servers to manage, no sync pipelines to maintain. Start in minutes.
Context is alive
A knowledge base shouldn't be a place where documentation goes to die. Topics in meridiaan are living context: updated, queryable, and immediately available to any agent.
One source of truth
Developers, managers, and AI agents all read from the same knowledge base. Decisions are recoverable. Context stays aligned across the whole team.
Who it's for
Independent developers
The primary user. Builds and maintains the knowledge base as the project evolves. Uses topics to document architectural choices, instructions, and key concepts, then queries them through an AI agent during daily work.
Managers
Not necessarily technical. Connects to the project's MCP server to ask questions about product status, recent progress, or feature details, in plain language, without reading through documentation.
Marketing and product
Uses the knowledge base to understand the product's characteristics, target audience, and positioning. Gets accurate answers fast, without bothering the engineering team.
Try it for free.
No credit card required. Set up your first project in minutes.