Zikkaron
Give Claude Code long-horizon memory across sessions so your agent recalls project context, decisions, and patterns instead of resetting every chat.
Overview
Zikkaron is a MCP server for the Build phase that provides biologically-inspired persistent memory for Claude Code MCP agents across sessions.
What is this MCP server?
- Biologically-inspired persistent memory engine aimed at Claude Code MCP workflows
- PyPI package zikkaron with stdio MCP transport (v1.0.0)
- Designed to survive session boundaries for coding agents
- GitHub-hosted open source (amanhij/Zikkaron)
- Fits local stdio MCP config alongside other dev tools
- MCP manifest version 1.0.0
- PyPI package identifier: zikkaron (stdio)
- Repository: github.com/amanhij/Zikkaron
Community signal: 60 GitHub stars.
What problem does it solve?
Coding agents forget prior context when sessions end, so you repeat architecture explanations and lose thread on multi-day solo builds.
Who is it for?
Solo developers on Claude Code running long-horizon builds who want local, agent-native memory without rolling a custom RAG pipeline first.
Skip if: Teams that need audited enterprise knowledge bases, multi-user ACLs, or memory with zero local Python dependencies.
What do I get? / Deliverables
After installing the stdio MCP server, your agent can read and write persistent memory so later sessions pick up project state instead of starting cold.
- Cross-session memory read/write tools for your agent
- Reduced re-onboarding overhead for the same repository
- Composable agent-tooling alongside other MCP servers
Recommended MCP Servers
Journey fit
Persistent agent memory is part of the build stack—same phase where you wire MCP servers, skills, and coding workflows. agent-tooling is the canonical shelf because Zikkaron is infrastructure for MCP agents, not a feature you ship to end users.
How it compares
Agent memory MCP server, not a hosted team docs product or a single-shot brainstorming skill.
Common Questions / FAQ
Who is Zikkaron for?
Indie and solo builders using Claude Code (or stdio MCP clients) who need durable agent memory across coding sessions.
When should I use Zikkaron?
Use it during build and agent-tooling setup when projects span multiple days and you want the agent to retain decisions, file context, and habits.
How do I add Zikkaron to my agent?
Install the PyPI package `zikkaron` (v1.0.0), add it as a stdio MCP server in Claude Code or your client’s MCP config, and point transport to the published identifier.