mm.tech
workflow stable Python

n8n Templates

Production n8n workflows with cross-session memory

Production-pattern hardened n8n workflow templates for voice agents, customer support, personal assistants. Multi-provider, powered by StudioMeyer Memory.

What it solves

Without persistent memory every automated conversation starts from zero. A voice agent forgets the caller. A support bot asks a returning customer for their email the third time. A personal assistant misses the project context you discussed yesterday. These are n8n workflows that fix that, and they are ready to import rather than something you assemble.

How it works

Every workflow follows the same loop: search the memory, reason with what came back, write the outcome back. The memory itself is StudioMeyer Memory, a hosted MCP backend with a knowledge graph, semantic search and tenant isolation. Each template is multi-provider from the start, so you pick OpenAI or Anthropic with one click and can add a Gemini or local Ollama branch with a single switch rule. The memory writes stay identical no matter which model you picked, which is the point: the workflow does not care.

When to use it

Take them when you already run n8n and want a working starting point for a voice agent, a support bot or a personal assistant that remembers people between calls. The free tier of the memory backend is activated with one click in the portal and needs no card.

When not to use it

They are templates, not a product. They assume a running n8n and someone who is comfortable editing a workflow. The memory-backed variants also assume a hosted backend, so if everything has to stay on your own hardware, start from the memory-free variant in the same repo instead.

Stars

★ 3

Forks

⑂ 1

Clones (14d)

⊟ 14

Language

Python

Updated

2026-08-10

Related repos