Blog
Engineering and product posts from the cachly team — persistent AI memory, causal_trace, brain_from_git, MCP server setup, and developer infrastructure.
Same Brain, Any Model: cachly Works with Claude, Cursor, Copilot, Windsurf and More
Your AI memory should not be locked to one vendor. cachly is the persistent, model-neutral memory layer — same Brain in Claude Code, Cursor, Windsurf, Copilot, Cline, Zed, and Continue. Bring your own model, keep your brain.
cachly is now on awesome-mcp-servers — 20,000 downloads and what comes next
PR #5174 merged. cachly is listed on punkpeye/awesome-mcp-servers, the most widely followed MCP discovery list. Nearly 20,000 npm downloads, Team Brain, CI-driven learning, gRPC for AI agents.
#1 on npm. One real user. Both true at the same time.
We're the most-installed MCP memory server on npm — and we have exactly one real user. Why neither number is failure, and the one promise we're making to every human. A manifesto.
We're the #1 MCP memory server on npm — 472 weekly installs in 2026
cachly's MCP server is now the most-installed memory server on npm: 472 weekly installs vs 17–27 for the next four competitors combined. Here's what worked, what didn't, and the lesson from our first real signups.
12,000 downloads in 23 days — MCP distribution & the memory pivot
How @cachly-dev/mcp-server reached 11,953 downloads in its first 23 days on npm — purely organic. What drove growth: the pivot from semantic cache to AI brain memory.
Zed Persistent Memory — Never Explain Your Stack Again
Zed forgets everything when a session ends. One command gives it permanent memory — auto-detects your setup, bootstraps from git history, shares across your team.
Windsurf Persistent Memory — Never Explain Your Stack Again
Windsurf forgets everything when a session ends. One command gives it permanent memory — auto-detects your setup, bootstraps from git history, shares across your team.
How to Give Cline Persistent Memory with MCP
Cline forgets everything between sessions. Add cachly's MCP server in 30 seconds to give Cline permanent memory — auto-learns from git history, shares across editors.
Claude Code Persistent Memory — Never Explain Your Stack Again
Claude Code forgets everything when a session ends. One command gives it permanent memory — auto-detects your setup, bootstraps from git history, shares across your team.
cachly vs CLAUDE.md — When a Static File Isn't Enough
CLAUDE.md is a good starting point. cachly is what you graduate to — automatic git learning, failure prediction, cross-editor sharing, and zero manual maintenance.
cachly vs mem0 — Which AI Memory is Right for Developers?
An honest comparison of cachly and mem0 for developer tooling. cachly wins on editor integration, git learning, causal analysis, and EU compliance. mem0 wins for Python LLM app builders.
How to Give GitHub Copilot Persistent Memory
GitHub Copilot forgets everything when you close VS Code. One command and 30 seconds fixes that permanently — auto-detects your editors, learns from your git history, shares across your team.
Introducing the Cognitive Cache: The Cache That Thinks
cachly v0.6 ships five capabilities no cache has ever had: causal root-cause tracing, knowledge decay scoring, brain diff, memory consolidation, and zero-config autopilot. This is not managed Redis. This is the world's first cache that thinks.
brain_from_git: Turn Your Git History into an AI Knowledge Base
Your commit log contains years of hard-won knowledge — bugs fixed, patterns discovered, gotchas avoided. brain_from_git reads it in 30 seconds and loads every lesson into your AI brain automatically.
Zero AI Onboarding: Full Codebase Context on Day One
New developers used to spend weeks learning what not to do. With brain_from_git and invite_link, your AI arrives pre-loaded with 2 years of team knowledge — before the first session ends.
Your AI Brain Now Speaks Arabic and Hebrew
cachly v0.5.48 adds native Arabic and Hebrew support: RTL tokenization, Arabic light stemming, 100+ stopwords, and bidirectional cross-language retrieval with the full synonym graph.
Memory Crystals: Distilling Team Knowledge into Instant AI Context
A Brain with 200 lessons is powerful but noisy. Memory Crystals distill everything your team has learned into a dense, always-fresh snapshot — injected into every AI session automatically.
Ambient Git Learning: Your Commit History as AI Knowledge
Your git log contains some of the richest operational knowledge in your codebase — and your AI has never read it. We built Ambient Git Learning to change that, with zero extra steps.
Team Telepathy: a Shared AI Brain for Engineering Teams
One developer fixes a hard bug. Five minutes later, every AI assistant on your team knows about it. No Slack message. No wiki update. No standup. That's Team Telepathy.
Store in Japanese, Recall in English — Cross-Language AI Memory
cachly v0.5.37 ships cross-language retrieval: store lessons in Japanese, Korean, Arabic, or Hebrew and recall them in English — no embeddings required. Built on a 130+ term synonym map.
Cachly SDK Integrations: 3 lines to semantic caching
Copy-paste examples for LangChain, Vercel AI SDK, OpenAI direct, Go, Ruby, PHP, Rust, and more. Every stack, same result: 60–90% fewer LLM API calls.
Free, Private Embeddings for Your AI Dev Brain — Powered by Ollama
We now run nomic-embed-text via Ollama on our infrastructure. No OpenAI API key needed. Your code never leaves Germany. Zero-cost semantic search for every Cachly Brain user.
Self-Host a Semantic LLM Cache in 5 Minutes
Run Cachly on your own server with one docker compose command. No Kubernetes, no cloud dependency, no data leaves your infra. Air-gapped, GDPR-ready, enterprise-grade.
How we cut LLM costs by 80% with Semantic Cache
Every user rephrases the same question differently. Without semantic caching you pay for each rephrasing. Here's how pgvector similarity search eliminates 60–90% of LLM API calls — with real numbers and 3 lines of code.
How I Built a VS Code Extension That Shows What My AI Learned
From `yo code` to a live status bar widget showing brain health, lesson count, and token savings — the full walkthrough including every gotcha. TypeScript, zero extra dependencies.
Building an IntelliJ Plugin in Kotlin: Status Bar + API
From build.gradle.kts to a live widget in IntelliJ IDEA, WebStorm, and all JetBrains IDEs — StatusBarWidgetFactory, PersistentStateComponent, Swing DialogWrapper, and the Gradle gotchas the docs don't mention.
See your AI Brain in VS Code and IntelliJ
New IDE plugins show brain health, lesson count, and recall stats directly in your status bar. VS Code and IntelliJ — zero config.
Your AI assistant never forgets — no embeddings required
We removed the #1 barrier to AI memory: the mandatory API key. Before: your assistant forgot everything. After: it remembers in 3ms. Zero config, works offline.
We built persistent memory for Claude Code
How we gave AI coding assistants a brain that survives across sessions — session briefings, lesson recall, team knowledge, and semantic search.
Search Your Japanese AI Brain in Romaji
cachly v0.5.37 adds Hepburn romaji search for katakana. Type 'kontena' and find コンテナ lessons instantly — digraphs, geminates, and long vowels all handled.
Your AI Never Forgets: Introducing Cachly Brain IDE Plugins
VS Code and IntelliJ plugins that show what your AI coding assistant has learned — lesson count, recall stats, and estimated token savings. One status bar widget, zero config.