
Basic Memory
Give your coding agent durable, local Markdown memory that syncs both ways so session context survives across projects.
Overview
Basic Memory is a MCP server for the Build phase that provides local-first Markdown knowledge management with bi-directional LLM sync for agent sessions.
What is this MCP server?
- Local-first knowledge store with bi-directional sync between LLM sessions and Markdown files
- Runs as stdio MCP via PyPI package basic-memory with uvx runtime hint
- You own plain Markdown on disk—readable outside any vendor UI
- Version 0.21.5 in catalog with explicit mcp subcommand in runtime arguments
- Fits long-horizon solo builds where decisions and specs must persist across weeks
- Package version 0.21.5; PyPI identifier basic-memory
- Stdio transport with uvx runtime hint and basic-memory mcp invocation
- Described as local-first with bi-directional LLM sync via Markdown
What problem does it solve?
Agent chats forget yesterday’s decisions, forcing solo builders to re-explain architecture, pricing, and scope every time they open a new session.
Who is it for?
One-person teams using Claude Code or Cursor who want offline-friendly, editor-visible memory without paying for a hosted knowledge SaaS.
Skip if: Multi-tenant team wikis with fine-grained ACLs, or builders who only need ephemeral chat with no long-term project ledger.
What do I get? / Deliverables
After registration, your agent reads and writes durable Markdown memory on your machine so project context stays aligned with files you control.
- On-disk Markdown corpus your agent can query and update
- Bi-directional sync between chat actions and files you edit manually
- Reusable project memory across agent sessions
Recommended MCP Servers
Journey fit
Persistent agent memory is built while you assemble tooling—before launch analytics—not as a one-off marketing task. Agent-tooling covers MCP servers that extend what your agent remembers and edits; Basic Memory is file-backed memory, not app backend code.
How it compares
Local Markdown memory MCP—not a cloud notebook app and not a RAG vector database server.
Common Questions / FAQ
Who is Basic Memory for?
It is for indie developers and agent power users who want persistent, file-based context that both humans and LLMs can update during long build cycles.
When should I use Basic Memory?
Use it while building agent-tooling—once you have recurring projects where specs, research, and implementation notes must survive across many sessions.
How do I add Basic Memory to my agent?
Install the basic-memory PyPI package, configure stdio MCP with uvx running basic-memory mcp per server metadata, then point your client at the local knowledge directory.