Part of the StudioMeyer MCP Stack — Built in Mallorca 🌴 · ⭐ if you use it
We have been building tools and systems for ourselves for the past two years. The fact that this repo is small and has few stars is not because it is new. It is because we only just decided to share what we have built. It is not a fresh experiment, it is a long story with a recent commit.
We love building things and sharing them. We do not love social media tactics, growth hacks, or chasing stars and followers. So this repo is small. The code is real, it gets used, issues get answered. Judge for yourself.
If it helps you, sharing, testing, and feedback help us. If it could be better, an issue is more useful. If you build something with it, tell us at hello@studiomeyer.io. That genuinely makes our day.
From a small studio in Palma de Mallorca.
n8n's built-in memory options (Postgres, Redis, Motorhead) handle short-term conversation history but stop there. StudioMeyer Memory is a long-term layer:
- Semantic search across past decisions, learnings, and sessions
- Knowledge graph with typed entities (people, projects, services) and relations
- Bi-temporal tracking so you can query "what did we know about X on date Y?"
- Confidence + decay so old contradicted facts fade automatically
- Multi-tenant on Supabase EU - DSGVO-compliant by default
This community node connects your n8n workflows to the StudioMeyer Memory MCP server (hosted SaaS at memory.studiomeyer.io) and exposes the most useful operations as a clean Resource/Operation tree. The custom Server URL field in the credential lets enterprise customers point at a managed-on-your-infra deployment after a commercial agreement, the public self-host bundle is a planned v0.2 release.
A voice agent that remembers callers across sessions. A customer support bot that knows the customer's preferences from three months ago. A personal assistant that builds a profile of you over years, not minutes. n8n is the perfect glue layer - but it lacks a real long-term memory primitive. This node fills that gap.
In your n8n instance: Settings → Community Nodes → Install
Enter the package name:
n8n-nodes-studiomeyer-memory
Or via npm if you self-host n8n:
npm install n8n-nodes-studiomeyer-memory- Get an API key. Sign in at studiomeyer.io/portal/login (Google, GitHub, Discord or email magic link). Inside the portal, click "Free Memory testen" - you get an API key with 200 free credits, no credit card. Copy the key from your dashboard at studiomeyer.io/portal/api.
- Add the credential in n8n. Settings → Credentials → New → "StudioMeyer Memory API". Paste your key.
- Drop the StudioMeyer Memory node into a workflow. Pick a Resource (Memory / Entity / Session / Insight) and an Operation.
The four core read/write operations.
| Operation | Maps to | Use case |
|---|---|---|
| Search | nex_search |
Semantic + keyword search across all memory types with temporal decay |
| Recall | nex_recall |
Faster lightweight recall (full-text only) |
| Learn | nex_learn |
Persist a fact, pattern, mistake, or workflow note |
| Decide | nex_decide |
Record a decision with rationale and confidence |
Knowledge graph operations.
| Operation | Maps to | Use case |
|---|---|---|
| Create | nex_entity_create |
Add a person, project, service, or concept |
| Observe | nex_entity_observe |
Append observations to an existing entity |
| Search | nex_entity_search |
Fuzzy search across names + observations |
| Relate | nex_entity_relate |
Connect entities (uses, depends_on, competes_with, custom) |
| Open | nex_entity_open |
Fetch one entity with all observations + relations |
Manage long-running agent sessions.
| Operation | Maps to | Use case |
|---|---|---|
| Start | nex_session_start |
Open a session, load context, get a session ID |
| End | nex_session_end |
Close a session and persist a summary |
| Recall Timeline | nex_recall_timeline |
Chronological view of recent activity |
Higher-level synthesis.
| Operation | Maps to | Use case |
|---|---|---|
| Synthesize | nex_synthesize |
Cluster recent learnings into an insight |
| Reflect | nex_reflect |
Surface emerging patterns + contradictions |
| Proactive Briefing | nex_proactive |
Status briefing - stale learnings, open decisions, knowledge gaps |
Two modes, same endpoint:
- API Key (default, recommended). Paste a key from your portal at studiomeyer.io/portal/api. First-time users sign in once at /portal/login, click "Free Memory testen", and the API key appears in the dashboard. 200 credits, no card.
- OAuth 2.1 Access Token. Pre-issued token from a PKCE flow against
memory.studiomeyer.io/authorize+/token. The OAuth discovery doc is at memory.studiomeyer.io/.well-known/oauth-authorization-server. The fully integrated browser flow (n8n OAuth2 credential type) ships in v0.2.
Both modes target https://memory.studiomeyer.io/mcp. The custom Server URL field is intended for managed-on-your-infra enterprise deployments (server source under commercial source-grant); a public self-host bundle is a planned v0.2 release.
Voice Webhook (Vapi/Retell)
→ Memory: Search (query: caller_phone, types: ["entity", "learning"])
→ AI Agent (with caller context in system prompt)
→ Memory: Learn (content: "Caller X asked about Y, agent answered Z")
WhatsApp Trigger
→ Entity: Search (query: customer_email, entityType: customer)
→ Entity: Open (entityRef: matched_id)
→ Anthropic Claude (with full customer context)
→ Memory: Learn (tag the ticket outcome)
Telegram Trigger
→ Session: Start (project: telegram-bot, agentId: user_id)
→ Memory: Search (query: telegram_message, recencyWeight: 0.5)
→ AI Agent
→ Memory: Learn (whatever the user just told you)
Free tier is 200 free credits activated by one click inside the portal at studiomeyer.io/portal/login (no card). Each operation (search, learn, entity-create, etc.) consumes one credit. Enough to evaluate the node and run a single bot in development. Pro is €29/month and lifts the limit to a generous monthly cap. Team is €49/month with multi-agent isolation. See studiomeyer.io/services/memory for the latest pricing and the per-tier cap.
Hosting model. This community node and its tests are MIT-licensed (use the node anywhere, commercial OK). The Memory server is currently hosted SaaS only (EU Frankfurt, Hetzner) - it is not self-hostable today. The server source lives in a private repo. If self-hosting is a hard requirement for your deployment (sovereignty, air-gap, customer contract), contact hello@studiomeyer.io about a commercial source-grant or managed-on-your-infra option.
- v0.1 (current, latest published v0.1.4): Memory / Entity / Session / Insight resources, API Key auth, OAuth 2.1 access token (paste a pre-issued token from a PKCE flow). The v0.1.1 - v0.1.4 patch line covers documentation hot-fixes (portal URL, free-tier credit count, signup flow) plus a robustness pass in v0.1.4 (errors carry the failing
itemIndex, richercontinueOnFailoutput, 97 unit tests). - v0.2: AI-Agent Memory Sub-Node (drops directly into the n8n AI Agent's Memory slot, replacing Postgres / Redis / Motorhead). Dedicated n8n OAuth2 credential type with full browser-based PKCE flow. Bundle
@modelcontextprotocol/sdkinto the published artifact to satisfy the n8n Verified Community Nodes "Zero Runtime Dependencies" requirement (deadline was 1 May 2026 - the v0.1.x line is shipped as an Unverified community node; Verified-status submission ships with v0.2 once SDK bundling lands and is smoke-tested). - v0.3: Streaming support for long search results, batch operations
- v0.4: Workflow templates published to n8n.io marketplace
git clone https://github.com/studiomeyer-io/n8n-nodes-studiomeyer-memory
cd n8n-nodes-studiomeyer-memory
npm install
npm run build
npm testTo test locally in your n8n instance:
npm link
cd ~/.n8n/custom
npm link n8n-nodes-studiomeyer-memory
n8n startPull requests welcome. Please open an issue first for non-trivial changes.
StudioMeyer is an AI and design studio based in Palma de Mallorca, working with clients worldwide. We build custom websites and AI infrastructure for small and medium businesses. Production stack on Claude Agent SDK, MCP and n8n, with Sentry, Langfuse and LangGraph for observability and an in-house guard layer.
MIT - see LICENSE.