cachly is a persistent AI memory platform for developers. It gives AI coding assistants like Claude Code, Cursor, GitHub Copilot and Windsurf a brain that remembers every lesson, fix, and architecture decision — forever. It connects via the MCP (Model Context Protocol) standard and includes 121 tools. Free tier available. Runs on German (EU) servers.

How does cachly work?

Run 'npx @cachly-dev/mcp-server@latest autopilot' once. The wizard auto-detects every AI editor you have installed (Claude Code, Cursor, Copilot, Windsurf, Cline, Zed) and writes the correct config for each. It then reads your entire git history with brain_from_git and loads years of team knowledge into your Brain before your first session. From that point, sessions start automatically, memory is shared across all your editors simultaneously, and a git post-commit hook teaches cachly from every commit.

Does cachly auto-detect my editors?

Yes. The cachly setup wizard automatically detects Claude Code, Cursor, GitHub Copilot, Windsurf, Cline, Zed, and Continue.dev — any editor that supports MCP. It writes the correct config file for each editor in one pass. You never manually edit JSON config files.

Is memory shared across all my AI editors?

Yes. cachly uses a single Brain that all your AI editors connect to simultaneously. A lesson remembered in Claude Code is instantly available in Cursor and GitHub Copilot. If your team uses different editors, all of you share the same persistent memory pool.

What is brain_from_git?

brain_from_git is a cachly tool that reads your entire git history before your first session and extracts lessons from every commit, PR, and revert. Your AI arrives knowing years of architectural decisions, bug fixes, and team conventions — without you writing a single line of documentation. Zero onboarding.

What is causal_trace?

causal_trace is a cachly tool that traces the history of any file or bug across your entire git history in seconds — replacing 30+ minutes of manual git blame. Describe a problem in plain English. It returns the root cause, the failure chain, and the exact fix that worked — with date, command, and file path.

What is brain_predict?

brain_predict is a cachly tool that scans your Brain for failure patterns before every deploy, migration, or dependency upgrade. It returns probability-weighted warnings based on your team's actual incident history — so you catch the next incident before it happens.

Does cachly work with Claude Code, Cursor, and GitHub Copilot?

Yes. cachly works with Claude Code, Cursor, GitHub Copilot, Windsurf, Cline, Zed, and Continue.dev — anywhere that supports MCP. Run 'npx @cachly-dev/mcp-server@latest autopilot' to configure all editors in one step. Memory is shared across all editors simultaneously.

Can cachly search memory across languages?

Yes. cachly uses semantic vector embeddings, not keyword search. A lesson stored in German appears when you search in English. A fix documented in Arabic matches a Japanese query about the same bug pattern. Supported languages include English, German, French, Spanish, Italian, Portuguese, Japanese, Chinese (Simplified and Traditional), Korean, Arabic, Hebrew, and more.

How is cachly different from mem0?

mem0 is a memory layer for Python LLM apps and chatbots — great for building AI products. cachly is built specifically for developer tooling: it connects to your AI editor via MCP, learns from your git history automatically, predicts failures before deploy, and gives your whole team shared memory. cachly runs on EU servers and is GDPR-native. For developers using Claude Code, Cursor, or Copilot, cachly is the right choice.

Is cachly GDPR compliant?

Yes. cachly runs exclusively on German servers (Hetzner). All data stays in the EU. No data is shared with third parties. cachly is fully GDPR compliant. An AVV (Auftragsverarbeitungsvertrag / Data Processing Agreement) is available for Business and Enterprise customers.

Store in Japanese, Recall in English — Cross-Language AI Memory

The Use Case

International engineering teams often have a language problem. The Japanese team stores lessons in Japanese. The Korean team writes in Korean. The shared AI memory is siloed — not because the Brain doesn't know these languages, but because keyword search can't bridge language boundaries. Until now.

How It Works — Without Embeddings

Semantic search with embeddings can find cross-lingual matches, but it requires an OpenAI API key, costs money per lookup, and adds latency. Most Brain users are on the free tier — no API key, no embeddings.

Our approach: a curated technical term synonym map, built directly into the tokenizer. When any document is indexed or any query is tokenized, every recognized technical term gets expanded to its equivalents in all supported languages:

Token: "deploy"
Expanded to: デプロイ (JA), 部署 (ZH), 배포 (KO), نشر (AR), פריסה (HE)
+ bigrams of all CJK variants added to the token stream

Token: デプロイ
Expanded to: "deploy", "deployment"
+ romaji "depuroi" added from katakana converter

The expansion happens at tokenize time — both when indexing (documents get synonym tokens) and when searching (queries get synonym tokens). This creates a shared token space across all 6 language pairs.

The Synonym Map — 130+ Terms

The map covers the technical vocabulary that actually appears in developer Brain entries:

Concept	JA	ZH	KO	AR	HE
deploy	デプロイ	部署	배포	نشر	פריסה
container	コンテナ	容器	컨테이너	حاوية	מיכל
server	サーバー	服务器	서버	سيرفر	שרת
error	エラー	错误	오류	خطأ	שגיאה
auth	認証	认证	인증	مصادقة	אימות
monitor	モニター	监控	모니터링	مراقبة	ניטור

Plus 100+ more: cache, database, build, test, install, log, port, cluster, debug, migration, and more.

Zero-Embedding, Zero-Cost

The entire cross-language lookup is a Map.get() call — O(1), no API, no network, no cost.

Traditional semantic search: 200–400ms latency, $0.0004/1000 tokens
Cross-lingual synonym lookup: <0.01ms latency, $0

This runs on the free tier, on every smart_recall call, in every Brain session.

Real-World Example

Team setup: Korean backend team stores lessons in Korean. English-speaking DevOps engineers query the Brain in English.

# Korean engineer stores a lesson
learn_from_attempts:
  topic: "deploy:k8s"
  outcome: "success"
  whatWorked: "배포 실패 원인: 포트 3000이 방화벽에 의해 차단됨. 포트를 열어 해결"

# English DevOps searches the next day
smart_recall("deployment failure port blocked")
→ ✅ Returns the Korean lesson, ranked first

# Symmetric: English lesson found by Korean query
learn_from_attempts:
  topic: "fix:redis"
  outcome: "success"
  whatWorked: "Redis connection timeout fixed: set timeout to 5000ms in config"

smart_recall("레디스 연결 오류")  # Korean: "Redis connection error"
→ ✅ Returns the English lesson

Supported Languages

The synonym graph now covers 6 language families: Japanese (hiragana + katakana), Chinese (simplified), Korean (hangul), Arabic (MSA technical vocabulary), Hebrew, and English. All language pairs are bidirectional.

Cross-language search activates automatically — no configuration, no flag, no API key. If you store in Japanese and recall in English, it just works.

Upgrade

npx @cachly-dev/mcp-server@latest autopilot