n8n-nodes-ejentum

An n8n community node that calls the Ejentum API and returns a task-matched cognitive operation (procedure + topology DAG + cognitive payload) for injection into the LLM step that follows.

Use the node before the AI Agent step on complex, multi-step, or multi-constraint tasks where the downstream model's default reasoning template would miss a constraint, take a shortcut, or drift across turns. Each call returns a cognitive operation: a structured procedure (numbered steps with a failure pattern to refuse and a falsification test) paired with an executable reasoning topology (a DAG of those steps with decision gates, parallel branches, bounded loops, and meta-cognitive exit nodes). The downstream model reads both layers from {{ $json.injection }} in its system prompt before producing its response.

One node, eight operations: four dynamic (Reasoning, Code, Anti-Deception, Memory) available on all tiers including the 30-day free trial, and four adaptive (Adaptive Reasoning, Adaptive Code, Adaptive Anti-Deception, Adaptive Memory) that run an additional adapter LLM step. Adaptive operations require the Go or Super tier.

Install

Follow the n8n community nodes installation guide:

1. Settings → Community Nodes → Install.
2. Enter n8n-nodes-ejentum.
3. Confirm the community-nodes risk acknowledgement.

The Ejentum node appears in the node picker.

Credentials

Create new Ejentum API credentials in n8n:

| Field | Value | |---|---| | API Key | your key (starts with ej_) | | API Base URL | default https://api.ejentum.com/harness/ unless self-hosting |

The Test button issues a real POST against the gateway and confirms the key is valid.

Get a key at ejentum.com/pricing.

Operations

Dynamic (all tiers)

| Operation | Mode string | Library size | |---|---|---:| | Reasoning | reasoning | 311 | | Code | code | 128 | | Anti-Deception | anti-deception | 139 | | Memory | memory | 101 |

Adaptive (Go or Super tier)

| Operation | Mode string | |---|---| | Adaptive Reasoning | adaptive-reasoning | | Adaptive Code | adaptive-code | | Adaptive Anti-Deception | adaptive-anti-deception | | Adaptive Memory | adaptive-memory |

Inputs and outputs

Input parameter (all operations)

• Query (string, required): 1-2 sentences describing the task the downstream LLM is about to perform. The harness uses this for retrieval.

Output format

The node has two output modes selected via a parameter:

| Output mode | Returned JSON | |---|---| | Injection String (default) | { "mode": "<mode>", "query": "<query>", "injection": "<text>" } | | Full Response | The full upstream payload: [ { "<mode>": "<text>" } ] |

In the default mode, drop {{ $json.injection }} into the next LLM node's system prompt.

Wire contract

The node issues:

POST https://api.ejentum.com/harness/
Headers: Authorization: Bearer <key>, Content-Type: application/json
Body:    { "query": <string>, "mode": <one of 8 mode strings> }
Response (200): [ { "<mode>": "<injection string>" } ]
Response (401|403|429): { "error": "..." }

Full wire contract, field structure of an injection, DAG syntax (token vocabulary used in the topology block), and a canonical dynamic-vs-adaptive comparison on the same query are documented in the ejentum-mcp README. The format is identical across this node and every Ejentum framework shim.

Canonical wiring

Ejentum is an agentic tool, not a pre-processing step. Wire it to the AI Agent node's tool input and let the agent call it on demand, alongside your other tools:

Chat Trigger ─► AI Agent ─────────────────► (response)
                  ├─ Chat Model: OpenAI / Anthropic / ...
                  ├─ Memory (optional)
                  └─ Tools
                       ├─ Ejentum (Reasoning / Code / Anti-Deception / Memory, + adaptive)
                       ├─ HTTP Request, Calculator, Vector Store, ...

The node is usableAsTool: true, so the AI Agent calls whichever Ejentum operation matches the sub-task it is reasoning about, mid-loop, the same way it calls any other tool. The returned injection (procedure, topology, suppression vectors, falsification test) shapes how the agent reasons before it acts. There is no manual prompt wiring: the agent decides when a task needs the harness.

Alternative for non-agent pipelines. In a deterministic flow with no AI Agent node, call Ejentum directly and inject its output into a downstream Chat Model's system prompt:

Trigger ─► Ejentum (query="{{ $json.task }}") ─► Chat Model
           system prompt:
           [REASONING CONTEXT]
           {{ $json.injection }}
           [END REASONING CONTEXT]

Place a fresh Ejentum call before each new sub-task; the injection's effect is strongest at the start of a branch.

ejentum-mcp alternative

n8n also ships an MCP Client node. The hosted MCP server at https://api.ejentum.com/mcp exposes the same eight tools with Bearer auth via your EJENTUM_API_KEY. Use this n8n node for tight n8n integration (credential vault, expression-friendly outputs, AI Agent routing); use the MCP route if you are already wiring multiple MCP servers into one workflow.

Compatibility

• n8n 2.9.x and above
• Node.js >= 20.15 on the n8n host
• Tested via eslint-plugin-n8n-nodes-base community + prepublish rule sets

Resources

• Homepage: https://ejentum.com
• Pricing: https://ejentum.com/pricing
• API reference: https://ejentum.com/docs/api_reference
• n8n community nodes: https://docs.n8n.io/integrations/community-nodes/

License

MIT