Aidenix

v0.3.0

Last updated Jul 27, 2026

Aidenix node for n8n — score a contact or a whole list against your ICP and write personalized outreach, check whether addresses are still worth contacting, profile the person and the company behind the lead, and manage ICP profiles, quota and opt-outs.

35 Weekly Downloads

338 Monthly Downloads

View on NPM GitHub Repository

Included Nodes

Aidenix

Description

n8n-nodes-aidenix

![npm version](https://www.npmjs.com/package/n8n-nodes-aidenix)
![npm downloads](https://www.npmjs.com/package/n8n-nodes-aidenix)
![License: MIT](./LICENSE.md)

An n8n community node for Aidenix — score each lead against your ICP and write a personalized first email and LinkedIn message, in your brand’s voice, from real research, in a single workflow step.

Drop in a LinkedIn slug or an email address. Aidenix researches the person across many data sources, scores fit against your ICP, and writes a first message that sounds like you wrote it for that one person — not a template with a name swapped in.

!Aidenix node — input parameters and the flat output it produces

—

Why Aidenix
Features
Prerequisites
Installation
Credentials
Quick start
Operations
Response schema
Common workflow patterns
Error handling
Idempotency & retries
Troubleshooting
Compatibility
Resources & support
License

—

Why Aidenix

Real personalization is the hard part of outbound, and a single merge tag doesn’t deliver it. A name and a title dropped into a template reads as mass-produced, and prospects ignore it.

Aidenix does the whole job in one node: it researches each lead across many data sources, scores fit against your ICP, and writes the first message in your brand’s voice — composed from your own positioning and the lead’s real context, not a snippet pasted into a template. You keep your sending tool and your strategy; Aidenix produces the message.

Input: jane@acme.com or jane-doe-12345
Output: fit score, person/company context, subject line, email body, LinkedIn DM, plus reasoning and Do/Don’t guidance

Open and licensed sources only — no traded PII, safe under GDPR and CCPA.

Use it when you want genuinely personal first touches at scale, without building the research-and-writing pipeline yourself.

Features

Business Fit — pass a LinkedIn slug or email; get an ICP fit score (0–100), a researched person summary, and a ready-to-send email + LinkedIn message written in your brand’s voice.
Email Intel — check an address before you spend an analysis on it: does the mailbox accept mail, and is the person still there.
Person Signals — the career moment behind the lead: months in the seat, what they publish themselves, when to reach them. Works even when the address is dead.
Company Signals — the account moment: hiring velocity from career dates, funding and product signals, and a growth / reorg / stealth / stable verdict with the timing.
Lists — assess addresses in bulk, score a whole batch of contacts, poll its status, and read the results page by page.
Account & ICP — quota and plan, the saved ICP profiles, building a profile from a website, and recording an opt-out.
Built-in idempotency — deterministic Idempotency-Key per item, so retried executions reuse the cached server response instead of paying for a duplicate AI run.
Automatic retry on 409 in_progress (still computing) and 504 timeout, with configurable backoff.
Standard n8n error handling via NodeApiError, with full Continue On Fail support.
HTTPS-only by default, with a runtime warning if a non-loopback Base URL is configured over plain HTTP.

Prerequisites

An n8n instance (Cloud or self-hosted), n8n v1.0+ recommended.
An Aidenix account and an API token — see Getting an API token.
For self-hosted n8n: Node.js 18.10+ (this matches n8n’s own engine requirement).

Installation

n8n Cloud / self-hosted UI

1. Open Settings → Community Nodes.
2. Click Install a community node.
3. Enter the package name:

   n8n-nodes-aidenix

4. Accept the risk prompt and click Install.

The Aidenix node appears in the node picker as Aidenix.

Manual install (self-hosted, advanced)

cd ~/.n8n/custom   # create if missing
npm install n8n-nodes-aidenix
restart n8n

Credentials

Getting an API token

1. Sign up or log in at aidenix.com.
2. Open your Profile → API Tokens.
3. Click Generate token, copy the value (it’s shown only once — store it somewhere safe).
4. In n8n, create a new Aidenix API credential and paste the token into the API Token field.

Credential fields

| Field | Required | Default | Description |
| ———- | ——– | ———————— | ——————————————————————————————— |
| API Token | Yes | — | Token issued in the Aidenix dashboard. Stored encrypted by n8n; sent as X-API-Token header. |
| Base URL | No | https://api.aidenix.com | Override only for dedicated tenants or self-hosted Aidenix. |

> Security note: the node refuses to log the token and warns if Base URL is plain HTTP to a non-loopback host. Keep Base URL on https:// in all real environments.

Quick start

A minimal “score-and-send” workflow:

1. Add a Schedule Trigger (or any trigger producing leads).
2. Add a Google Sheets / Airtable node to read a list of leads — each row should have an email (or linkedin_slug) column.
3. Add the Aidenix node:
– Operation: Business Fit
– Query: ={{ $json.email }}
– Leave Options at their defaults.
4. Add an IF node:
– Condition: ={{ $json.fitscorepct }} ≥ 70
5. On the true branch, add Gmail / Outlook with:
– Subject: ={{ $json.email_subject }}
– Body: ={{ $json.email_message }}

That’s the whole pipeline. Idempotency is on by default — re-running a failed execution reuses the cached Aidenix response instead of paying for it twice. Running the workflow again on the same lead in a new execution is a fresh computation.

Operations

Business Fit

Evaluates ICP fit for a contact and generates personalized outreach.

| Parameter | Type | Required | Description |
| ——— | —— | ——– | —————————————————————————- |
| Query | string | Yes | LinkedIn slug (jane-doe-12345) or email (jane@acme.com). Auto-detected. |

#### Options

| Option | Default | Description |
| ———————— | ————– | —————————————————————————————————————————————— |
| Idempotency Key Strategy | Deterministic | Deterministic — UUID v5 from workflowId + executionId + itemIndex + query. Replays reuse the cached response. Random — new UUID v4 per item. Custom — supply your own. |
| Idempotency Key | — | Only used when strategy is Custom. Useful when the key comes from an upstream node, e.g. ={{ $json.idempotency_key }}. |
| Max Retries (409 / 504) | 10 | Retries when the API responds with 409 in_progress or 504 timeout. |
| Retry Delay (Ms) | 5000 | Delay between retries. |

The retry options apply to every operation. The idempotency options apply to Business Fit only — every other operation either reads (a repeat is safe by nature) or is idempotent on the server side.

Email Intel

Is the address worth contacting at all. Two independent questions — does the mailbox accept mail, and is the person still there — because a live mailbox at a stale address bounces the deal, not the message. Run it before Business Fit to avoid personalizing for someone who left.

Branch on verdict (current-likely, former-person-moved, former-flagged-historic, domain-rebrand, personal-mailbox, ambiguous-generic-alias, unknown-not-in-data, uncertain) or, more simply, on recommendation.action (send, caution, verify, skip).

Person Signals

Who the person is and what is going on with them: role and grade, the career trajectory from dates, what they publish themselves, whose content they read, and a verdict with the career moment and how to approach it. It works even when the address is dead, because the person is resolved to their current employer.

| Parameter | Type | Required | Description |
| ——— | —— | ——– | ———————————————————————— |
| Person | string | Yes | Email, LinkedIn profile URL, or slug. |

Asked by email, the response carries no name and no slug — turning a mailbox into an identity is exactly what this API refuses to do, and identifierswithheld says so. Ask by profile URL or slug and the name comes back, since you already hold it. When matchconfidence is low, check linked_profiles: above one, several people stand behind that address and the most corroborated was taken.

Company Signals

The account moment behind the lead: firmographics and revenue, team makeup, hiring velocity from career dates (which catches layoffs a static card misses), activity signals, geo footprint, and a growth / reorg / stealth / stable verdict with the outreach timing. Employee names are never returned — only business signals. One call per account covers every lead from it.

Email Intel (Bulk)

The same address check across a whole list in one call, so a list can be triaged before any analysis is spent on it. The slow layers apply per address, so a long list with Deliverability or Enrich on takes minutes.

Score a List

Submits a batch of contacts and returns a batch ID. Contacts analysed earlier come from cache and cost nothing new; duplicates inside the list are dropped before anything is spent.

Batch Status

How far along a batch is: done, pending, failed, served from cache, and how many scored 60 or above — the list worth exporting. Poll this rather than the results page.

| Parameter | Type | Required | Description |
| ——— | —— | ——– | ————————————— |
| Batch ID | string | Yes | The ID returned by Score a List. |

Batch Results

The analyses themselves, one page at a time. Feed next_cursor from the previous page back into Cursor; an empty cursor means you are at the end.

Account Context

Plan, quota used and left, when the period resets, and which ICP profile scores are currently weighted against. Worth checking before a large batch.

ICP Profiles

Every saved profile with its ID, and which one is active.

Build ICP Profile

Reads a website and saves it as an ICP profile. The new profile becomes the active one, so every later score is weighted by it — put this at the start of a workflow deliberately, not in a loop.

| Parameter | Type | Required | Description |
| ———– | —— | ——– | ———————————- |
| Website URL | string | Yes | The site to read. |

Record Opt-Out

Records that a contact asked to be left alone. An email, a phone number, or a LinkedIn URL — recognised and stored in normalised form. Recording the same contact twice does not create a second record.

| Parameter | Type | Required | Description |
| ——— | —— | ——– | ———————————————- |
| Contact | string | Yes | Email, phone number, or LinkedIn URL. |

Response schema

The node returns a flat JSON object per input item:

{
  "job_id": "9f1c…",
  "status": "completed",
  "query": "jane-doe-12345",
  "fitscorepct": 82,
  "person_description": "Jane — CTO at a fintech startup, ~50 engineers, recently raised Series B.",
  "email_subject": "Quick question about your payment flow",
  "email_message": "Hi Jane, …",
  "linkedin_message": "Hi Jane, …",
  "logic_explanation": [
    "Role matches ICP (technical decision-maker)",
    "Company at the right stage (post-Series B)"
  ],
  "strategy": "Practical, problem-first approach",
  "strategy_do": ["Mention the specific pain", "Be concrete about ROI"],
  "strategy_dont": ["Don't open with generic claims", "Don't reference unrelated case studies"]
}

All fields are safe to reference directly in downstream nodes via ={{ $json.field_name }}.

Common workflow patterns

1. Trigger → Score → Conditional send

!Workflow: Execute → Aidenix → IF → Gmail / No-op

Trigger ─▶ Aidenix (Business Fit) ─▶ IF (fitscorepct ≥ 70) ─▶ Gmail (send)
                                                                  └▶ No-op (skip)

Score every lead, send only to those above your threshold, drop the rest. Swap the manual trigger for a Schedule or Webhook in production.

2. Webhook → Real-time enrichment for CRM

Webhook (CRM "new lead") ─▶ Aidenix ─▶ HTTP Request (PATCH /crm/leads/{{id}})

Push fit score and personalized copy back into HubSpot / Pipedrive in seconds.

3. LinkedIn slug enrichment from form fills

Typeform Trigger ─▶ Aidenix (Query = ={{ $json.linkedin_url | extractSlug }}) ─▶ Slack (#sales)

Notify sales the moment a high-fit lead fills out a form.

Error handling

The node maps Aidenix’s HTTP responses into n8n’s standard error flow:

Enable Continue On Fail on the node to push errored items into the success branch as { error, query, idempotencyKey } — useful for partial-success batch workflows.

Idempotency & retries

The Aidenix API uses Idempotency-Key to deduplicate expensive AI runs. By default this node generates a deterministic key from workflowId:executionId:itemIndex:query, which means:

Re-running a failed execution reuses the previously-computed result for free.
Running the same workflow on the same lead in a new execution generates a new key — you’ll get a fresh computation. This is usually what you want, since context (your ICP definition) may have changed.

To force a single deterministic key across executions (e.g. “score this lead once per day, regardless of how many times the workflow fires”), switch the strategy to Custom and supply your own:

={{ $json.email + ':' + $now.format('yyyy-MM-dd') }}

Troubleshooting

The node returns 409 in_progress even after all retries.
Increase Max Retries or Retry Delay in the node options. Large batches can occasionally exceed the default 10×5s window during peak load.

402 quota_exhausted.
Your plan has run out of credits — top up in the Aidenix dashboard. The node will surface this immediately rather than silently degrading.

401 unauthorized after rotating the token.
n8n credentials cache. Re-open the credential in n8n, paste the new token, save.

The score looks wrong / outreach feels generic.
Open a ticket at aidenix.com — fit scoring is tied to your account’s ICP definition, which is configured outside n8n.

Base URL warning in execution logs.
You’ve set Base URL to a non-loopback http:// host. Switch to https:// — the warning means the API token is being sent in cleartext.

Compatibility

n8n: tested on 1.x. Uses the v1 INodeType API.
Node.js: 18.10+ (matches n8n’s engines requirement).
Aidenix API: this package is maintained alongside the API by the Aidenix team — versions are kept compatible.

Resources & support

Aidenix — aidenix.com
API docs — aidenix.com/api
n8n community nodes guide — docs.n8n.io/integrations/community-nodes
Issues & feature requests — GitHub issues
Email — info@aidenix.com