Integration

Long-term memory for Zed

Zed's AI panel is fast and minimal, and it starts every conversation with no memory of the last one. Add Engram as a remote context server in Zed's settings and the panel gains six durable memory tools alongside its built-in code awareness. There's no extension to install, just a settings.json edit.

Install

Three steps: sign up for an Engram API key, paste a BYOK LLM-provider key on /models, then drop the snippet below into Zed.

Three steps to memory in your agent

  1. Sign up. Free, no card. You'll land on a Getting Started page that walks the next two steps.
  2. Add your LLM key. Engram is BYOK. Paste an OpenAI / Anthropic / Groq / Together / Fireworks key and we'll route every extraction and query call through your provider. You pay your provider directly. We never see your inference.
  3. Paste the snippet below into your agent and restart it. Use Authorization: Bearer <api-key>with the API key from your portal.

engram-zed: Zed editor AI panel

Configured as a remote context_server directly in Zed's settings.json. No extension install required. OAuth supported too: drop the Authorization header and Zed will prompt for OAuth on first use. Source: github.com/lumetra-io/engram-zed.

  1. Open ~/.config/zed/settings.json (or +,Open Settings) and add:
  2. ~/.config/zed/settings.json
    {
      "context_servers": {
        "engram": {
          "url": "https://mcp.lumetra.io/mcp/sse",
          "headers": {
            "Authorization": "Bearer <api-key>"
          }
        }
      }
    }
  3. Open Zed's AI panel → context servers list. engram should appear with the six tools listed on hover.

What you can do once memory's wired in

  • Recall this morning's refactor plan when you come back to the codebase after lunch
  • Remember why you chose a particular crate or library so Zed doesn't suggest swapping it out
  • Pull up the conventions you've explained for Rust vs TS in a multi-language repo
  • Skip re-explaining the legacy module's quirks every time Zed touches it

FAQ

Do I need to install a Zed extension?

No. Engram is configured as a remote `context_server` directly in `settings.json`, and Zed treats it as a first-class MCP source with no extension involved.

OAuth or Bearer header?

Both work. Bearer is simpler: paste the key in `Authorization`. OAuth is cleaner if you share the Zed config across multiple developers, since each person authenticates with their own Lumetra account on first use.

How do I confirm it's loaded?

Open the AI panel and look at the context servers list. `engram` should appear with the six tools listed. Try the prompt `"Use engram store_memory to remember 'zed verification'"` to sanity-check the tool call path.

Ship durable memory in Zed today

Free tier: 10K memories and 50K retrievals per month. No credit card. Same Engram backend powers all 41 integrations, so memories you write from one client are immediately queryable from the rest.