Description
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 flowsOpenAI API Key: optional, sent asx-openai-keywhen providedAnthropic API Key: optional, sent asx-anthropic-keywhen providedGoogle API Key: optional, sent asx-google-keywhen provided- You can run some flows with an empty Mnexium key (trial provisioning).
- Provider keys are still required when the selected model/provider requires them.
User Messageis the primary input.- If empty, the node can fall back to incoming fields such as
message,text,input, orprompt. - Optional Mnexium controls include
subjectid,chatid,learn,history,recall,log, and summarization/prompt policy options. - Advanced records controls are available for
mnx.records: 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/testwithout writing cache.Sync Cache: execute/api/v1/integrations/:id/syncand persist mapped values.Send Webhook: sign and send payloads to/api/v1/integrations/:id/webhook.- Provide either
Webhook Signaturedirectly or aWebhook Secretso the node can compute the HMAC header. - If
Timestampis empty, the node sends the current Unix timestamp. - Optional
Event IDsupports backend deduplication. - HTTP Method
- Path
- Headers (JSON)
- Query Params (JSON)
- Body (JSON) for non-GET methods
- Absolute URLs are restricted to HTTPS on
mnexium.comandwww.mnexium.com. - Relative paths are resolved against
https://www.mnexium.com. 401 Unauthorized:- No model response:
scan:n8nreturns404:
Notes:
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| Memory | List, Search, Create, Get, Update, Delete, List Superseded, Restore, List Recall Events, Get Claims For Memory |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:
– 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:
Webhook notes:
Custom API Call
Custom API call is available only at Resource: Custom Request.
Inputs:
Security:
Troubleshooting
– Check Mnexium key and provider keys.
– Verify backend deployment/state for your Mnexium environment.
– Confirm provider key matches the chosen model provider.
– Publish package to npm first, then run scanner.
Development
npm install
npm run build
npm run lint
Cloud review helper:
npm run scan:n8n