cachly← All posts
Tutorial · May 14, 2026

Zed Persistent Memory — Never Explain Your Stack Again

Zed is one of the fastest code editors available, with native AI assistant support via Claude. But like every AI tool, it forgets everything between sessions. This guide shows you how to give Zed permanent memory in 30 seconds using cachly's MCP server.

Why Zed forgets — and why that costs you

Zed's AI assistant operates within a context window. When you close the editor, switch projects, or start a new session, everything is gone. No memory of:

  • ✗Your architectural decisions and the reasoning behind them
  • ✗Bugs you spent hours fixing last week
  • ✗Your team's naming conventions and code style decisions
  • ✗Which third-party libraries have known issues in your stack
  • ✗What the codebase looked like before the last major refactor

Every session is a clean slate. You spend the first minutes re-explaining context instead of writing code.

The fix: cachly MCP server

cachly connects to Zed via MCP (Model Context Protocol), which Zed supports natively as "context servers." Once connected, Zed's AI can use cachly's tools for permanent, cross-session recall.

smart_recall

Finds stored lessons and context by meaning, not by exact key

causal_trace

Follows a problem back to its root cause through your stored lessons

brain_predict

Returns likely failures and the fixes that worked before

brain_from_git

Turns your last 100 commits into lessons during setup

learn_from_attempts

Stores a lesson after a fix, so the next session finds it

recall_best_solution

Returns the best known solution for a topic, with a confidence marker

Setup: one command, 30 seconds

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

The wizard auto-detects Zed and writes the correct MCP config to ~/.config/zed/settings.json. One browser click to sign in. No API keys to manage:

// ~/.config/zed/settings.json (or .zed/settings.json in project root)
{
  "context_servers": {
    "cachly": {
      "command": {
        "path": "npx",
        "args": ["-y", "@cachly-dev/mcp-server@latest"],
        "env": {
          "CACHLY_JWT": "your-token-here"
        }
      }
    }
  }
}

Zed uses context_servers instead of mcpServers — the setup wizard writes the right format automatically. If you also use Claude Code, Cursor, or Windsurf, the wizard configures all of them in the same pass.

brain_from_git: context before session one

After setup, cachly automatically runs brain_from_git — it reads your last 100 commits and loads the lessons into your Brain:

# What brain_from_git extracts from your git log:
✓ Fix: JWT expiry check before signature validation (commit a3f91c2)
✓ Revert: Redis pub/sub removed — race condition at scale (commit 8b2e441)
✓ Decision: tRPC over REST — type safety across monorepo (commit f9c1203)
✓ Warning: Stripe webhook order not guaranteed — always idempotent

Zed arrives at your first session already knowing years of your team's decisions — without you writing a single documentation line.

What a session looks like with cachly

Open Zed. On the first AI assistant call, smart_recall fires automatically:

> Good morning. Here's what I remember:
>
> Last session: payments refactor — stripe-webhook.ts
> Team learned: always set idempotency_key for Stripe charges
>   → commit a3f91c2, 2025-11-14
>
> brain_predict: ⚠️ Moderate deploy risk
> The k8s namespace change in PR #441 matches a past
> race condition pattern (incident 2024-08-12).

No prompting. No manual context. Zed just knows.

cachly works across all your editors

One Brain, shared across every tool in your workflow:

Zed

✓ supported

Claude Code

✓ supported

Cursor

✓ supported

Windsurf

✓ supported

Run the setup once — cachly configures all detected editors automatically. Switch between editors without losing any context.

Give Zed memory in 30 seconds

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

Auto-detects Zed, Claude Code, Cursor, Windsurf, Copilot. Free tier, no credit card, German servers.

Get your free token →

cachly.dev

For solo developers, teams and whole companies

Explain it once. Every AI knows it.

cachly keeps what you explain to your AI in every session. Claude Code, Cursor, Copilot and Windsurf read your rules, decisions and solved bugs: for you alone, for your whole team and for every team in your company. Servers in Germany, with SSO and a DPA for companies.

Servers in the EU · one command, restart your editor, done · 1 project free

Four ways to install

Claude Code
/plugin marketplace add cachly-dev/cachly-mcp
/plugin install cachly-brain@cachly

Then turn on auto-update once: /plugin → Marketplaces → cachly → Enable auto-update

Any MCP editor
npx @cachly-dev/mcp-server@latest autopilot