ahmetakyurt/zipmem-mcp
by Various
Local-first MCP server + CLI giving terminal AI agents an endless, low-token memory via Anchored Compacting
MCP
ahmetakyurt/zipmem-mcp
Added 13 July 2026
Overview
A local-first MCP server and CLI that provides terminal-based AI agents with a low-token memory system via Anchored Compacting. It stores compressed memory snapshots locally, reducing token usage for long-running agent interactions.
Best for
Best for
Developers building local-first terminal AI agents who need low-token memory persistence.
Use cases
- Persisting agent context across terminal sessions
- Reducing token costs in AI-assisted development workflows
- Enabling local memory for offline-capable CLI agents
How to use
Install
npm install -g zipmem-mcp Tools exposed
zipmem_load_memoryzipmem_checkpointzipmem_save_and_compact
Tested with
Claude Code
Example client config
{\n "mcpServers": {\n "zipmem-mcp": {\n "command": "npx",\n "args": ["zipmem-mcp"]\n }\n }\n} Notes
A local-first MCP server and CLI that provides terminal-based AI agents with a low-token memory system via Anchored Compacting. It stores compressed memory snapshots locally, reducing token usage for long-running agent interactions.
0 stars on GitHub. Last updated 2026-07-01. Licensed MIT.
Use cases
- Persisting agent context across terminal sessions
- Reducing token costs in AI-assisted development workflows
- Enabling local memory for offline-capable CLI agents
Pros
- Saves tokens by compressing and archiving memory locally
- Works offline with no cloud dependency
- Lightweight TypeScript implementation
Cons
- Limited community adoption (zero stars and no documentation)
- Unproven reliability of Anchored Compacting algorithm
- Requires manual CLI integration into existing agent setups
Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.
Pros
- Saves tokens by compressing and archiving memory locally
- Works offline with no cloud dependency
- Lightweight TypeScript implementation
Cons
- Limited community adoption (zero stars and no documentation)
- Unproven reliability of Anchored Compacting algorithm
- Requires manual CLI integration into existing agent setups
Open-source & AI alternatives
Swap-in tools that solve the same job. Weigh the trade-offs before you commit.
Get the free Developer’s Field Guide
A 27-page field guide to the AI coding workflow with Claude. Claude Code, MCP servers, the prompt patterns that work, and what to delegate. Free.
Enter your work email. We send it straight over, plus a few short notes worth knowing. Unsubscribe any time.
