mnexium

Official Mnexium community node for n8n

Package Information

Downloads: 116 weekly / 885 monthly
Latest Version: 0.1.11
Author: Mnexium

Available Nodes

Mnexium icon
Mnexium
Use Mnexium memory, claims, profiles, state, records, integrations, prompts, and audit APIs

Documentation

n8n-nodes-mnexium

Official Mnexium community node for n8n.

Use Mnexium to add persistent memory, structured claims, profiles, state, records, prompts, and audit visibility to your AI workflows.

Docs: https://www.mnexium.com/docs

Installation

Install in your n8n environment:

npm install n8n-nodes-mnexium

Restart n8n after installation.

Credentials

Credential type: Mnexium API + Model Keys

  • Mnexium API Key: optional for supported trial/free-tier flows
  • OpenAI API Key: optional, sent as x-openai-key when provided
  • Anthropic API Key: optional, sent as x-anthropic-key when provided
  • Google API Key: optional, sent as x-google-key when provided

Notes:

  • You can run some flows with an empty Mnexium key (trial provisioning).
  • Provider keys are still required when the selected model/provider requires them.

Supported Resources And Operations

Resource Operations
Chat Message a Model
Chat History List Chats, Read Chat, Delete Chat
Memory List, Search, Create, Get, Update, Delete, List Superseded, Restore, List Recall Events, Get Claims For Memory
Claim Get Truth, Get Slot Value, Get History, List Slots, Get Graph, Create, Get, Retract
Profile Get Profile, Update Fields, Delete Field, Get Schema
State Get, Set, Delete
Record Schema List Schemas, Create/Update Schema, Get Schema
Record List, Create, Get, Update, Delete, Query, Search
System Prompt List, Create, Get, Update, Delete, Resolve
Memory Policy List, Create, Get, Update, Delete, Resolve
Integration List, Create, Get, Update, Delete, Test Pull, Sync Cache, Send Webhook
Audit List
Custom Request Custom API Call

Quick Start

  1. Add a Mnexium node to a workflow.
  2. Select Resource: Chat and Operation: Message a Model.
  3. Set Model (for example gpt-4.1-mini).
  4. Set User Message directly or pass text from incoming JSON.
  5. Execute and inspect response data.

Chat Input Behavior

For Chat -> Message a Model:

  • User Message is the primary input.
  • If empty, the node can fall back to incoming fields such as message, text, input, or prompt.
  • Optional Mnexium controls include subject_id, chat_id, learn, history, recall, log, and summarization/prompt policy options.
  • Advanced records controls are available for mnx.records:
    • Records Learn Mode: off | auto | force
    • Records Tables (JSON Array): allowlist for extraction targets
    • Records Sync: wait for write completion before response
    • Records Recall: inject relevant records into context

Integrations

Use Resource: Integration when you want typed access to Mnexium inbound connectors instead of building raw custom requests.

Supported integration operations:

  • List: fetch project integrations, with optional inactive rows.
  • Create: define pull/webhook connectors with templates, output mappings, auth, and webhook secrets.
  • Get / Update / Delete: inspect and manage existing connector configs.
  • Test Pull: execute /api/v1/integrations/:id/test without writing cache.
  • Sync Cache: execute /api/v1/integrations/:id/sync and persist mapped values.
  • Send Webhook: sign and send payloads to /api/v1/integrations/:id/webhook.

Webhook notes:

  • Provide either Webhook Signature directly or a Webhook Secret so the node can compute the HMAC header.
  • If Timestamp is empty, the node sends the current Unix timestamp.
  • Optional Event ID supports backend deduplication.

Custom API Call

Custom API call is available only at Resource: Custom Request.

Inputs:

  • HTTP Method
  • Path
  • Headers (JSON)
  • Query Params (JSON)
  • Body (JSON) for non-GET methods

Security:

  • Absolute URLs are restricted to HTTPS on mnexium.com and www.mnexium.com.
  • Relative paths are resolved against https://www.mnexium.com.

Troubleshooting

  • 401 Unauthorized:
    • Check Mnexium key and provider keys.
    • Verify backend deployment/state for your Mnexium environment.
  • No model response:
    • Confirm provider key matches the chosen model provider.
  • scan:n8n returns 404:
    • Publish package to npm first, then run scanner.

Development

npm install
npm run build
npm run lint

Cloud review helper:

npm run scan:n8n

Links

Discussion