Adding Persistent Memory to Claude Code with claude-mem —…
    Neura MarketNeura Market/Stable Diffusion
    ChatGPTChatGPTClaudeClaudeGeminiGeminiCursorCursorGrokGrokPerplexityPerplexityStable DiffusionStable Diffusion
    DeepSeekDeepSeekCoPilotCoPilotMidjourneyMidjourney
    View All Directories
    OverviewPromptsBlogVideosGuidesCoursesCommunityModelsLoRAsComfyUI WorkflowsTrending
    Stable DiffusionBlogAdding Persistent Memory to Claude Code with claude-mem — Plus a DIY Lightweight Alternative
    Back to Blog
    Adding Persistent Memory to Claude Code with claude-mem — Plus a DIY Lightweight Alternative
    claudecode

    Adding Persistent Memory to Claude Code with claude-mem — Plus a DIY Lightweight Alternative

    kanta13jp1 April 13, 2026
    0 views

    The Problem: Claude Code Forgets Everything Every time you start a new Claude Code...

    The Problem: Claude Code Forgets Everything

    Every time you start a new Claude Code session, the slate is wiped clean. Your coding style preferences, project architecture decisions, yesterday's debugging session — all gone.

    You end up repeating yourself: "We use Supabase, not Firebase. The Edge Functions are in supabase/functions/. Don't use dummy data."

    claude-mem fixes this by adding persistent memory across sessions. It hit 46K GitHub stars within 48 hours of launch. I installed it, built a lightweight DIY alternative first, and here's what I found.

    What is claude-mem?

    GitHub: https://github.com/thedotmack/claude-mem

    A plugin that gives Claude Code a long-term memory. It automatically captures what you do during sessions and injects relevant context into future conversations.

    Architecture

    • 5 Lifecycle Hooks: SessionStart / UserPromptSubmit / PostToolUse / Stop / SessionEnd
    • SQLite + Chroma: Hybrid search (keyword + vector similarity)
    • Bun HTTP Worker: Background service on localhost:37777
    • MCP Tools: 3-layer progressive disclosure (search → timeline → get_observations)
    • Web UI: Visual memory browser

    Installation

    npx claude-mem install
    npx claude-mem start  # Requires Bun
    

    The DIY Alternative I Built First

    Before discovering claude-mem, I built a minimal memory system using just two PowerShell scripts and Claude Code's native hooks API.

    PostToolUse Hook (auto-capture.ps1)

    Triggered after every Bash or Write tool use. Captures git commits and new file creations to a daily markdown file:

    memory/auto-capture/2026-04-13.md
    - 09:15 [abc1234] feat: Add user authentication
    - 09:32 [Write] auth_middleware.dart
    - 10:01 [def5678] fix: Token refresh logic
    

    SessionStart Hook (session-resume.ps1)

    Reads the last 3 days of captures and injects them as context when a new session starts. The AI immediately knows what you've been working on.

    Registration in settings.json

    {
      "hooks": {
        "PostToolUse": [{
          "matcher": "Bash|Write",
          "hooks": [{
            "type": "command",
            "command": "powershell -File auto-capture.ps1"
          }]
        }],
        "SessionStart": [{
          "hooks": [{
            "type": "command",
            "command": "powershell -File session-resume.ps1"
          }]
        }]
      }
    }
    

    Head-to-Head Comparison

    Featureclaude-memDIY Hooks
    Setupnpx install (1 command)2 scripts, manual
    Auto-captureAll tool usagegit commits + Write only
    SearchVector similarity + keywordgrep (text search)
    Web UIlocalhost:37777None
    DependenciesBun + SQLite + (Chroma)None
    Token costLLM compression (Gemini = free)Zero
    Git-friendlyDB file (gitignored)Markdown files (shareable)
    Multi-instanceSession-scoped isolationFile sharing for coordination

    Running Both Together

    The good news: they coexist perfectly. claude-mem registers as a plugin, DIY hooks register directly in settings.json. Both fire on the same events without conflict.

    When claude-mem shines

    • Smart compression: Uses an LLM (Gemini/Claude) to summarize tool outputs into compact observations
    • Semantic search: "What did I do with the auth system last week?" actually works
    • Web dashboard: Visual overview of what's been captured

    When DIY hooks shine

    • Zero dependencies: No server, no database, no runtime
    • Team sharing: Markdown files can be committed to git and shared across instances
    • Full control: You decide exactly what gets captured and how
    • Truly free: No API calls whatsoever

    Cost Optimization Tip

    claude-mem defaults to using Claude API for compression, which consumes your tokens. Switch to Gemini (free) to eliminate this:

    // ~/.claude-mem/settings.json
    {
      "CLAUDE_MEM_PROVIDER": "gemini",
      "CLAUDE_MEM_GEMINI_API_KEY": "your-free-key-from-aistudio.google.com"
    }
    

    Our 3-Layer Memory Architecture

    In our project (Flutter Web + Supabase, 3 parallel Claude Code instances), we use a layered approach:

    LayerToolPurpose
    L1: Intra-sessionclaude-mem (SQLite)Auto-record all tool usage, semantic search
    L2: Inter-sessionDIY hooks (markdown)Git commit history, cross-instance sharing
    L3: Cross-projectNotebookLM Master BrainDeep research, long-term architectural knowledge

    Verdict

    claude-mem delivers on its promise of turning Claude Code from a "disposable tool" into a "growing partner." The vector search and Web UI are genuinely useful features that are hard to replicate with simple scripts.

    However, for teams that want zero dependencies, zero token cost, and git-friendly memory sharing, a DIY hook approach is a solid starting point.

    My recommendation: Start with DIY hooks for minimal memory, then layer on claude-mem when you need semantic search and automatic compression.


    Built with Claude Code | Project: https://my-web-app-b67f4.web.app/ #ClaudeCode #AI #buildinpublic

    Tags

    claudecodeaiproductivityllm

    Comments

    More Blog

    View all
    Context bankruptcy: The case for strategic forgetting for AI Agentsai

    Context bankruptcy: The case for strategic forgetting for AI Agents

    Most of us have seen a coding agent fail to complete a task we know it can do. We just don't...

    J
    James O'Reilly
    Parallel Compliance Engine: Drive-to-Sheets Multi-Agent Orchestrationgooglecloud

    Parallel Compliance Engine: Drive-to-Sheets Multi-Agent Orchestration

    When building Generative AI applications, developers often encounter a massive bottleneck: sequential...

    A
    Aryan Irani
    Is It Ethical to Post and Ask About Circuits on Dev.to?discuss

    Is It Ethical to Post and Ask About Circuits on Dev.to?

    I’ve been thinking about sharing some electronic circuit posts on Dev.to — small circuits, DIY...

    C
    codebunny20
    The One-Click Exporter: AI Studio Antigravity, Probed to Its Limitsagents

    The One-Click Exporter: AI Studio Antigravity, Probed to Its Limits

    What nobody tells you about exporting your multi-agent prototype to a local workspace. Every...

    L
    leslysandra
    Guarding the till while autonomous data agents do the diggingagenticarchitect

    Guarding the till while autonomous data agents do the digging

    Autonomous agents are genuinely good at answering messy business questions. Give one an LLM and a set...

    S
    Sireesha Pulipati
    Return on Attention: Why AI Code Reviews Are Wearing Us Outai

    Return on Attention: Why AI Code Reviews Are Wearing Us Out

    PR volume went up, ticket quality didn't, and the gap got filled with LLMs on both sides of the review: bots reviewing, bots replying, bots occasionally arguing with bots about priorities that only existed in a teammate's head. Our CEO named the actual problem, and it's bigger than code review.

    C
    christine

    Stay up to date

    Get the latest Stable Diffusion prompts, rules, and resources delivered to your inbox weekly.

    Neura Market LogoNeura Market

    Discover the best AI prompts, plugins, and resources for Stable Diffusion and more.

    Content Types

    • Rules
    • Prompts
    • MCPs
    • Agents
    • Guides

    Platforms

    • ChatGPT Directory
    • Claude Directory
    • Gemini Directory
    • Cursor Directory
    • Grok Directory
    • Perplexity Directory
    • DeepSeek Directory
    • CoPilot Directory
    • Stable Diffusion Directory
    • Midjourney Directory
    • All Directories

    Resources

    • Blog
    • Documentation
    • Help Center
    • Marketplace

    Legal

    • Privacy Policy
    • Terms of Service

    © 2026 Neura Market. All rights reserved.

    |

    Not affiliated with any AI platform vendors.

    Ready-made automations for this

    Workflows from the Neura Market marketplace related to this Stable Diffusion resource

    • Persistent Chat Memory with GPT-4o-mini & Qdrant Vector DBn8n · $24.99 · Related topic
    • Unlimited Website Down Checker - Uptime Robot Alternativen8n · $14.99 · Related topic
    • Node-Fetch Lightweight HTTP Client for Node.jsn8n · $6.99 · Related topic
    • Debug: Lightweight Node.js & Browser Debugging Utilityn8n · $6.99 · Related topic
    Browse all workflows