Zikkaron
A persistent memory layer for Claude that stores conversation history, decisions, and project context in SQLite. Exposes 24 MCP tools including `remember` to store outcomes, `recall` to retrieve past sessions by project name, and `get_project_context` for directory-specific history. Built on neuroscience-inspired algorithms like Hopfield energy scoring and spreading activation for retrieval without runtime LLM calls. Benchmarked at 86.8% Recall@10 on LoCoMo and 0.404 on BEAM's 10M token conversations. Handles context window compaction through hippocampal replay to preserve decisions that would otherwise get summarized away. Install via pip, add one JSON block to your MCP config, and drop three lines in your global CLAUDE.md to automate recall on every session start.
- GitHub stars
- 60
- tools
- 24
- transport
- STDIO
- last updated
- Mar 6, 2026
Tools
24 toolsPublic tool metadata for what this MCP can expose to an agent.
rememberStore a memory through the predictive coding write gate
Store a memory through the predictive coding write gate
No parameter schema in public metadata yet.
recallMulti-signal retrieval with heat-weighted ranking
Multi-signal retrieval with heat-weighted ranking
No parameter schema in public metadata yet.
forgetDelete a memory
Delete a memory
No parameter schema in public metadata yet.
validate_memoryCheck staleness against current file state
Check staleness against current file state
No parameter schema in public metadata yet.
get_project_contextGet hot memories for a directory
Get hot memories for a directory
No parameter schema in public metadata yet.
consolidate_nowForce a consolidation cycle
Force a consolidation cycle
No parameter schema in public metadata yet.
memory_statsSystem statistics across all subsystems
System statistics across all subsystems
No parameter schema in public metadata yet.
rate_memoryUsefulness feedback for metamemory tracking
Usefulness feedback for metamemory tracking
No parameter schema in public metadata yet.
recall_hierarchicalQuery the fractal hierarchy at a specific abstraction level
Query the fractal hierarchy at a specific abstraction level
No parameter schema in public metadata yet.
drill_downNavigate into a memory cluster
Navigate into a memory cluster
No parameter schema in public metadata yet.
create_triggerSet prospective triggers that fire on matching context
Set prospective triggers that fire on matching context
No parameter schema in public metadata yet.
get_project_storyAutobiographical narrative of a project
Autobiographical narrative of a project
No parameter schema in public metadata yet.
add_ruleNeuro-symbolic constraints for filtering and re-ranking
Neuro-symbolic constraints for filtering and re-ranking
No parameter schema in public metadata yet.
get_rulesList active rules
List active rules
No parameter schema in public metadata yet.
navigate_memoryTraverse concept space via successor representations
Traverse concept space via successor representations
No parameter schema in public metadata yet.
get_causal_chainCausal ancestors and descendants for an entity
Causal ancestors and descendants for an entity
No parameter schema in public metadata yet.
assess_coverageEvaluate knowledge coverage with gap identification
Evaluate knowledge coverage with gap identification
No parameter schema in public metadata yet.
detect_gapsFind isolated entities, stale regions, missing connections
Find isolated entities, stale regions, missing connections
No parameter schema in public metadata yet.
checkpointSnapshot working state for compaction recovery
Snapshot working state for compaction recovery
No parameter schema in public metadata yet.
restoreReconstruct context after compaction via Hippocampal Replay
Reconstruct context after compaction via Hippocampal Replay
No parameter schema in public metadata yet.
anchorMark critical facts as compaction-resistant
Mark critical facts as compaction-resistant
No parameter schema in public metadata yet.
install_hooksEnable auto-capture, context injection, and compaction recovery hooks
Enable auto-capture, context injection, and compaction recovery hooks
No parameter schema in public metadata yet.
sync_instructionsUpdate CLAUDE.md with latest Zikkaron capabilities
Update CLAUDE.md with latest Zikkaron capabilities
No parameter schema in public metadata yet.
seed_projectBootstrap memory for an existing project in one scan
Bootstrap memory for an existing project in one scan
No parameter schema in public metadata yet.
Sponsors
Agent CaffeineSponsoredClose your lid, come back to finished work
SnugSponsoredThe clipboard manager for the AI eraPromote your product hereReach 125k+ visitors monthly
WaxSponsoredBlock anything with supervision, without erasing your iPhonePromote your product hereReach 125k+ visitors monthly
knobsSponsoredSee your site on any phone, any day, in any languageReadme
Zikkaron
Zikkaron (זיכרון) is Hebrew for "memory."
Your AI forgets you every time you close the tab. Every architecture decision you explained, every debugging rabbit hole you went down together, every "remember, we're using Postgres not SQLite" correction. Gone. You start the next session a stranger to your own tools.
Zikkaron is a persistent memory engine for Claude Code built on computational neuroscience. It remembers what you worked on, how you think, what you decided and why. Not as a dumb text dump that gets shoved into context, but as a living memory system that consolidates, forgets intelligently, and reconstructs the right context at the right time.
26 subsystems. 24 MCP tools. Runs entirely on your machine. One SQLite file.
Two minutes to never repeat yourself again
Details
- Repository
- github.com/amanhij/zikkaron
- Category
- AI & LLM Tools
- Package
zikkaron- Transport
- STDIO
- Registry
- MCP registry, active
Related AI & LLM Tools MCP servers
View all
Akbdnotitia/akbGit-backed org memory for AI agents — hybrid search, tables, files & a URI graph over MCP.52stars
Mastercard Developers MCPmastercard/developers-agent-toolkitAccess Mastercard Developers APIs, documentation, and API specifications through MCP.0stars
Palinodephasespace-labs/palinodeInspectable, correctable project memory in git-versioned Markdown for AI agents.46stars
NeuroStackraphasouthall/neurostackToken-efficient MCP memory for Markdown vaults. Tiered search, GraphRAG, AI memories.43stars
Mcp Probe Kitmybolide/mcp-probe-kitSpec-driven dev workflow MCP: start_* orchestration, memory, GitNexus, quality gates.41stars