creation

2026-04-19 09:05:01 +00:00 · 2026-03-15 12:01:17 +08:00
parent 2377275d7d
commit f091dabbc3
1 changed files with 808 additions and 0 deletions
--- a/GPT-4o-mini-13514/readme-13514.md
+++ b/GPT-4o-mini-13514/readme-13514.md
@@ -0,0 +1,808 @@
+Find and add lookalike influencers to HubSpot with influencers.club and GPT-4o-mini
+
+https://n8nworkflows.xyz/workflows/find-and-add-lookalike-influencers-to-hubspot-with-influencers-club-and-gpt-4o-mini-13514
+
+
+# Find and add lookalike influencers to HubSpot with influencers.club and GPT-4o-mini
+
+## 1. Workflow Overview
+
+This workflow listens for new HubSpot contacts, checks whether the new contact is a creator, finds similar creators through the influencers.club API, enriches each lookalike profile, analyzes them with GPT-4o-mini, and then creates or updates those lookalikes in HubSpot.
+
+Its main use case is creator discovery for influencer marketing teams: starting from an existing high-performing creator in HubSpot, the workflow identifies comparable creators and pushes enriched, AI-qualified prospects back into the CRM.
+
+### 1.1 Trigger and HubSpot Contact Retrieval
+The workflow starts from a HubSpot trigger on contact creation, then fetches the full contact record by ID.
+
+### 1.2 Seed Creator Preparation
+It extracts contact fields, enriches the person by email through influencers.club, normalizes the returned creator data, and determines whether the contact is actually a creator.
+
+### 1.3 Lookalike Discovery
+If the contact is a creator, the workflow calls the influencers.club lookalike endpoint using the detected main platform and account handle.
+
+### 1.4 Lookalike Enrichment and Normalization
+Each returned lookalike is processed one at a time, enriched again for full profile data, and transformed into a normalized profile object.
+
+### 1.5 AI Qualification and HubSpot Upsert
+GPT-4o-mini produces a structured evaluation of the lookalike creator, and the workflow then sends the analyzed result into HubSpot for create/update handling.
+
+### 1.6 Error Handling and Silent Drops
+If the original enrichment fails, the workflow stops with an explicit error. Non-creators are dropped silently. Failed lookalike enrichments are also configured to continue without breaking the loop.
+
+---
+
+## 2. Block-by-Block Analysis
+
+## 2.1 Workflow Documentation and Context Notes
+
+**Overview:**  
+This block contains all sticky notes used to describe the workflow, setup requirements, and per-stage behavior. These notes are not executable but are essential for maintainers and automation agents interpreting the workflow.
+
+**Nodes Involved:**  
+- 📋 OVERVIEW
+- Note: Trigger
+- Note: Get Contact
+- Note: Extract Email
+- Note: Enrich by Email
+- Note: Stop and Error
+- Note: Normalize Seed
+- Note: Filter
+- Note: Lookalike API
+- Note: Loop
+- Note: Enrich Lookalike1
+- Note: AI Agent
+- Note: HubSpot Out
+- Sticky Note1
+
+### Node Details
+
+#### 📋 OVERVIEW
+- **Type and role:** Sticky Note; high-level workflow description.
+- **Configuration choices:** Documents purpose, flow, platform detection behavior, and required credentials.
+- **Key expressions or variables used:** Mentions `main_platform`.
+- **Input/output connections:** None.
+- **Version-specific requirements:** None.
+- **Edge cases or failure types:** None; documentation only.
+- **Sub-workflow reference:** None.
+
+#### Note: Trigger
+- **Type and role:** Sticky Note; describes the HubSpot trigger behavior.
+- **Configuration choices:** Notes that the trigger fires on `contact.creation` and can be extended with `contact.propertyChange`.
+- **Input/output connections:** None.
+- **Version-specific requirements:** Depends on HubSpot app event support.
+- **Edge cases:** If the HubSpot app is not configured for the intended event, nothing will trigger.
+- **Sub-workflow reference:** None.
+
+#### Note: Get Contact
+- **Type and role:** Sticky Note; explains HubSpot contact retrieval.
+- **Configuration choices:** Mentions retrieved properties `firstname`, `company`, `hs_full_name_or_email`.
+- **Edge cases:** Warns that more properties may need to be added for custom CRM setups.
+
+#### Note: Extract Email
+- **Type and role:** Sticky Note; describes field mapping into `Email` and `Company Name`.
+
+#### Note: Enrich by Email
+- **Type and role:** Sticky Note; explains the email-based enrichment call.
+- **Edge cases:** Indicates failures go to Stop and Error.
+
+#### Note: Stop and Error
+- **Type and role:** Sticky Note; documents the hard-stop failure path.
+
+#### Note: Normalize Seed
+- **Type and role:** Sticky Note; explains the JavaScript normalization of the seed creator.
+- **Configuration choices:** Notes platform auto-detection, follower/engagement tiering, growth trend detection, and URL cleanup.
+
+#### Note: Filter
+- **Type and role:** Sticky Note; describes creator-only filtering.
+- **Configuration choices:** Documents true branch vs silent drop branch.
+
+#### Note: Lookalike API
+- **Type and role:** Sticky Note; explains the lookalike search and possible filters.
+- **Configuration choices:** Notes `paging.limit` and optional filtering dimensions like engagement and brand deals.
+
+#### Note: Loop
+- **Type and role:** Sticky Note; describes sequential item processing using looping.
+
+#### Note: Enrich Lookalike1
+- **Type and role:** Sticky Note; explains full enrichment per lookalike.
+- **Configuration choices:** Notes that errors are suppressed so the loop can continue.
+
+#### Note: AI Agent
+- **Type and role:** Sticky Note; explains GPT-4o-mini structured analysis output.
+- **Configuration choices:** Suggests adjusting the system prompt for campaign context.
+
+#### Note: HubSpot Out
+- **Type and role:** Sticky Note; describes final HubSpot upsert.
+- **Configuration choices:** Recommends mapping AI outputs into custom HubSpot properties.
+
+#### Sticky Note1
+- **Type and role:** Sticky Note; marketing/context note with external link.
+- **Configuration choices:** Includes link to influencers.club content.
+- **Context or link:** `https://influencers.club/creatorbook/find-similar-creators-to-top-performers/`
+
+---
+
+## 2.2 Trigger and Contact Retrieval
+
+**Overview:**  
+This block starts the workflow when a HubSpot contact is created and retrieves the full contact record using the emitted `contactId`.
+
+**Nodes Involved:**  
+- HubSpot Trigger
+- Get Contact by ID
+
+### Node Details
+
+#### HubSpot Trigger
+- **Type and technical role:** `n8n-nodes-base.hubspotTrigger`; event-driven entry point.
+- **Configuration choices:**  
+  - Uses a HubSpot Developer API credential.
+  - Configured with one event in `eventsUi`; based on the note, this is intended for `contact.creation`.
+  - `maxConcurrentRequests` is set to `5`.
+- **Key expressions or variables used:** Outputs `contactId`.
+- **Input and output connections:**  
+  - No input node; entry point.
+  - Outputs to **Get Contact by ID**.
+- **Version-specific requirements:** `typeVersion: 1`.
+- **Edge cases or potential failure types:**  
+  - HubSpot webhook subscription not configured correctly.
+  - Developer app missing required scopes.
+  - Event mismatch if the app is not subscribed to contact creation.
+  - Concurrency issues if HubSpot bursts exceed allowed processing rate.
+- **Sub-workflow reference:** None.
+
+#### Get Contact by ID
+- **Type and technical role:** `n8n-nodes-base.hubspot`; retrieves a HubSpot contact by ID.
+- **Configuration choices:**  
+  - Operation: `get`
+  - Authentication: HubSpot app token
+  - Contact ID comes from `={{ $json.contactId }}`
+  - Requests properties: `firstname`, `company`, `hs_full_name_or_email`
+- **Key expressions or variables used:**  
+  - `{{ $json.contactId }}`
+- **Input and output connections:**  
+  - Input from **HubSpot Trigger**
+  - Output to **Extract Email**
+- **Version-specific requirements:** `typeVersion: 2.2`
+- **Edge cases or potential failure types:**  
+  - Invalid or missing `contactId`
+  - HubSpot auth/token failure
+  - Property path differences depending on portal configuration
+  - The later Set node expects fields that may not actually be included here, especially email and associated company structures
+- **Sub-workflow reference:** None.
+
+---
+
+## 2.3 Seed Contact Extraction and Enrichment
+
+**Overview:**  
+This block maps HubSpot contact fields into simpler variables and sends the contact email to influencers.club for creator enrichment.
+
+**Nodes Involved:**  
+- Extract Email
+- Enrich by Email
+- Stop and Error
+
+### Node Details
+
+#### Extract Email
+- **Type and technical role:** `n8n-nodes-base.set`; creates a simplified payload.
+- **Configuration choices:**  
+  - Assigns `Email` from `={{ $json.properties.email.value }}`
+  - Assigns `Company Name` from `={{ $json["associated-company"].properties.name.value }}`
+- **Key expressions or variables used:**  
+  - `{{ $json.properties.email.value }}`
+  - `{{ $json["associated-company"].properties.name.value }}`
+- **Input and output connections:**  
+  - Input from **Get Contact by ID**
+  - Output to **Enrich by Email**
+- **Version-specific requirements:** `typeVersion: 3.4`
+- **Edge cases or potential failure types:**  
+  - Likely field mismatch: the previous HubSpot node does not explicitly request `email`.
+  - `associated-company` may not exist.
+  - Expression evaluation failure if nested properties are absent.
+- **Sub-workflow reference:** None.
+
+#### Enrich by Email
+- **Type and technical role:** `n8n-nodes-influencersclub.influencersClub`; creator enrichment by email.
+- **Configuration choices:**  
+  - Uses `email = {{ $json.Email }}`
+  - `onError` is set to `continueErrorOutput`, which allows the error output branch to be used
+- **Key expressions or variables used:**  
+  - `{{ $json.Email }}`
+- **Input and output connections:**  
+  - Input from **Extract Email**
+  - Main output 0 to **Normalize**
+  - Error output to **Stop and Error**
+- **Version-specific requirements:** `typeVersion: 1`
+- **Edge cases or potential failure types:**  
+  - Missing/invalid email
+  - influencers.club auth errors
+  - No creator found for the given email
+  - Rate limits or transient API failures
+- **Sub-workflow reference:** None.
+
+#### Stop and Error
+- **Type and technical role:** `n8n-nodes-base.stopAndError`; explicit execution termination.
+- **Configuration choices:**  
+  - Error message: `API enrichment failed - contact is not a creator or API error`
+- **Key expressions or variables used:** None.
+- **Input and output connections:**  
+  - Input from **Enrich by Email** error branch
+  - No outputs
+- **Version-specific requirements:** `typeVersion: 1`
+- **Edge cases or potential failure types:**  
+  - This intentionally stops execution; useful for observability but may be noisy if many contacts are non-creators.
+- **Sub-workflow reference:** None.
+
+---
+
+## 2.4 Seed Creator Normalization and Filtering
+
+**Overview:**  
+This block transforms the enrichment payload into a flat creator object, auto-detects the main platform, and filters out non-creators before lookalike search.
+
+**Nodes Involved:**  
+- Normalize
+- Filter - Is Creator?
+
+### Node Details
+
+#### Normalize
+- **Type and technical role:** `n8n-nodes-base.code`; JavaScript normalization for seed creator data.
+- **Configuration choices:**  
+  - Accepts either array-wrapped or direct payloads.
+  - Extracts `result` if present.
+  - Builds a platform map for TikTok, Instagram, YouTube, and Twitter.
+  - Detects `main_platform` by highest follower/subscriber count.
+  - Computes:
+    - `followers_fmt`
+    - `follower_tier`
+    - `engagement_tier`
+    - `growth_trend`
+  - Normalizes handle and strips URL query params from images.
+  - Outputs a flattened object suitable for filtering and lookalike lookup.
+- **Key expressions or variables used:**  
+  - Uses `$input.first().json[0] || $input.first().json`
+  - Produces keys such as:
+    - `email`
+    - `full_name`
+    - `handle`
+    - `main_platform`
+    - `profile_url`
+    - `followers`
+    - `engagement_percent`
+    - `is_creator`
+    - `platforms_available`
+- **Input and output connections:**  
+  - Input from **Enrich by Email**
+  - Output to **Filter - Is Creator?**
+- **Version-specific requirements:** `typeVersion: 2`
+- **Edge cases or potential failure types:**  
+  - Unknown or empty platform data leads to `main_platform = 'unknown'`
+  - Missing usernames cause `handle = null`
+  - Assumes certain enrichment fields exist for each platform
+  - YouTube URL construction depends on `custom_url`
+- **Sub-workflow reference:** None.
+
+#### Filter - Is Creator?
+- **Type and technical role:** `n8n-nodes-base.if`; boolean gate.
+- **Configuration choices:**  
+  - Condition: `{{ $json.is_creator }}` equals `true`
+  - Only the true branch is connected
+- **Key expressions or variables used:**  
+  - `{{ $json.is_creator }}`
+- **Input and output connections:**  
+  - Input from **Normalize**
+  - True output to **Find Similar Creators**
+  - False output unconnected, so non-creators are dropped silently
+- **Version-specific requirements:** `typeVersion: 2`
+- **Edge cases or potential failure types:**  
+  - If `is_creator` is null/undefined/false, the item disappears without error
+  - Strict boolean validation may reject non-boolean variants
+- **Sub-workflow reference:** None.
+
+---
+
+## 2.5 Lookalike Discovery
+
+**Overview:**  
+This block queries influencers.club for similar creators based on the seed creator’s detected platform and handle, then splits the returned accounts into individual items.
+
+**Nodes Involved:**  
+- Find Similar Creators
+- Split Out
+- Loop Over Items (Lookalike)
+
+### Node Details
+
+#### Find Similar Creators
+- **Type and technical role:** `n8n-nodes-influencersclub.influencersClub`; lookalike search.
+- **Configuration choices:**  
+  - Resource: `creator`
+  - Operation: `findLookalikes`
+  - Platform: `={{ $json.main_platform }}`
+  - Filter key: `username`
+  - Filter value: `={{ $json.handle }}`
+  - Filter groups for TikTok, Twitch, Twitter, YouTube, OnlyFans, Instagram, and advanced filters are present but empty
+- **Key expressions or variables used:**  
+  - `{{ $json.main_platform }}`
+  - `{{ $json.handle }}`
+- **Input and output connections:**  
+  - Input from **Filter - Is Creator?**
+  - Output to **Split Out**
+- **Version-specific requirements:** `typeVersion: 1`
+- **Edge cases or potential failure types:**  
+  - If `main_platform` is `unknown`, the API may fail or return no useful results
+  - Missing handle causes poor or failed matching
+  - API rate limit/auth failure
+  - Empty result set
+- **Sub-workflow reference:** None.
+
+#### Split Out
+- **Type and technical role:** `n8n-nodes-base.splitOut`; splits array field into separate items.
+- **Configuration choices:**  
+  - Field to split: `accounts`
+- **Key expressions or variables used:** None.
+- **Input and output connections:**  
+  - Input from **Find Similar Creators**
+  - Output to **Loop Over Items (Lookalike)**
+- **Version-specific requirements:** `typeVersion: 1`
+- **Edge cases or potential failure types:**  
+  - Missing `accounts` field
+  - `accounts` not being an array
+  - Empty array leading to no downstream processing
+- **Sub-workflow reference:** None.
+
+#### Loop Over Items (Lookalike)
+- **Type and technical role:** `n8n-nodes-base.splitInBatches`; sequential loop controller.
+- **Configuration choices:**  
+  - Default batching behavior; effectively processes one item at a time in this pattern
+- **Key expressions or variables used:** None.
+- **Input and output connections:**  
+  - Input from **Split Out**
+  - Output 1 to **Enrich by Handle (Full)** for the next item
+  - Output 0 is loop completion
+  - Receives a return connection from **AI Agent (Lookalike)** to continue the loop
+- **Version-specific requirements:** `typeVersion: 3`
+- **Edge cases or potential failure types:**  
+  - Improper loop-back connection could stall the workflow
+  - Large result sets may still create long-running executions
+- **Sub-workflow reference:** None.
+
+---
+
+## 2.6 Lookalike Enrichment and Profile Normalization
+
+**Overview:**  
+This block takes each lookalike candidate, fetches the full creator profile, and converts potentially different response formats into one standardized structure for AI analysis.
+
+**Nodes Involved:**  
+- Enrich by Handle (Full)
+- Normalize Profile (Lookalike)
+
+### Node Details
+
+#### Enrich by Handle (Full)
+- **Type and technical role:** `n8n-nodes-influencersclub.influencersClub`; full creator enrichment by handle.
+- **Configuration choices:**  
+  - Resource: `creator`
+  - Operation: `enrichByHandle`
+  - Handle: `={{ $json.profile.username }}`
+  - Platform: `={{ $('Normalize').item.json.main_platform }}`
+  - `onError` is `continueErrorOutput`
+- **Key expressions or variables used:**  
+  - `{{ $json.profile.username }}`
+  - `{{ $('Normalize').item.json.main_platform }}`
+- **Input and output connections:**  
+  - Input from **Loop Over Items (Lookalike)**
+  - Output to **Normalize Profile (Lookalike)**
+- **Version-specific requirements:** `typeVersion: 1`
+- **Edge cases or potential failure types:**  
+  - Missing `profile.username` in the lookalike payload
+  - Platform mismatch if lookalike results are not actually on the same platform as the seed
+  - API enrichment failure; configured not to abort execution
+- **Sub-workflow reference:** None.
+
+#### Normalize Profile (Lookalike)
+- **Type and technical role:** `n8n-nodes-base.code`; transforms lookalike profile data into a consistent schema.
+- **Configuration choices:**  
+  - Runs once for each item.
+  - Supports multiple source formats:
+    - Format A: enriched result object
+    - Format B: lightweight lookalike object with `profile`
+    - Format C: already-flat enrichment
+  - Creates nested `normalized_profile` with sections:
+    - `identity`
+    - `cross_platform_links`
+    - `contact`
+    - `audience`
+    - `follower_growth`
+    - `engagement`
+    - `content`
+    - `monetization`
+    - `network`
+    - `flags`
+    - `dedup_key`
+  - Includes helper logic for:
+    - URL cleanup
+    - handle normalization
+    - follower formatting
+    - growth trend calculation
+    - cross-platform link inference
+    - collaborator deduplication
+    - hashtag extraction
+- **Key expressions or variables used:**  
+  - Uses `item.json`
+  - Outputs `normalized_profile`
+- **Input and output connections:**  
+  - Input from **Enrich by Handle (Full)**
+  - Output to **AI Agent (Lookalike)**
+- **Version-specific requirements:** `typeVersion: 2`
+- **Edge cases or potential failure types:**  
+  - Assumes Instagram-centric normalization in several branches
+  - Some fields may be null if enrichment is partial
+  - If input is an array unexpectedly, it maps over entries, which may change item cardinality
+  - Deduplication key may be null when both handle and user ID are absent
+- **Sub-workflow reference:** None.
+
+---
+
+## 2.7 AI Analysis
+
+**Overview:**  
+This block sends each normalized lookalike profile to a LangChain AI agent backed by GPT-4o-mini and forces the output into a strict JSON schema.
+
+**Nodes Involved:**  
+- AI Agent (Lookalike)
+- OpenAI Model (Lookalike)1
+- Structured Output Parser (Lookalike)1
+
+### Node Details
+
+#### OpenAI Model (Lookalike)1
+- **Type and technical role:** `@n8n/n8n-nodes-langchain.lmChatOpenAi`; language model provider.
+- **Configuration choices:**  
+  - Model: `gpt-4o-mini`
+  - No extra options set
+- **Key expressions or variables used:** None.
+- **Input and output connections:**  
+  - Connected as AI language model to **AI Agent (Lookalike)**
+- **Version-specific requirements:** `typeVersion: 1.3`
+- **Edge cases or potential failure types:**  
+  - Invalid OpenAI credentials
+  - Token/rate limit issues
+  - Model availability changes
+- **Sub-workflow reference:** None.
+
+#### Structured Output Parser (Lookalike)1
+- **Type and technical role:** `@n8n/n8n-nodes-langchain.outputParserStructured`; validates/parses model output against a manual JSON schema.
+- **Configuration choices:**  
+  - Manual schema with required fields:
+    - `summary`
+    - `strengths`
+    - `concerns`
+    - `audience_analysis`
+    - `growth_analysis`
+    - `content_analysis`
+    - `brand_fit`
+    - `flags`
+    - `recommended_action`
+    - `recommended_action_reason`
+    - `estimated_reach_per_post`
+  - Includes enums for engagement quality, trajectory, niche clarity, and recommended action
+- **Key expressions or variables used:** None.
+- **Input and output connections:**  
+  - Connected as AI output parser to **AI Agent (Lookalike)**
+- **Version-specific requirements:** `typeVersion: 1.3`
+- **Edge cases or potential failure types:**  
+  - Model output may fail schema validation
+  - Numeric fields may be returned as strings by the model
+- **Sub-workflow reference:** None.
+
+#### AI Agent (Lookalike)
+- **Type and technical role:** `@n8n/n8n-nodes-langchain.agent`; prompts the model and produces structured analysis.
+- **Configuration choices:**  
+  - Prompt type: defined manually
+  - Uses `{{ JSON.stringify($json.normalized_profile, null, 2) }}`
+  - Strong system instruction to return valid JSON only
+  - Output parser enabled
+- **Key expressions or variables used:**  
+  - `{{ JSON.stringify($json.normalized_profile, null, 2) }}`
+- **Input and output connections:**  
+  - Main input from **Normalize Profile (Lookalike)**
+  - AI model input from **OpenAI Model (Lookalike)1**
+  - AI output parser input from **Structured Output Parser (Lookalike)1**
+  - Main outputs to:
+    - **HubSpot — Lookalike Creator**
+    - **Loop Over Items (Lookalike)** to continue iteration
+- **Version-specific requirements:** `typeVersion: 3.1`
+- **Edge cases or potential failure types:**  
+  - Schema mismatch despite parser
+  - Hallucinated values if profile data is sparse
+  - Long prompts if enriched profile payload becomes large
+  - Depending on agent behavior, output may contain only parsed JSON and not the original profile fields, which matters for the next HubSpot node
+- **Sub-workflow reference:** None.
+
+---
+
+## 2.8 HubSpot Output
+
+**Overview:**  
+This block is intended to create or update the analyzed lookalike in HubSpot.
+
+**Nodes Involved:**  
+- HubSpot — Lookalike Creator
+
+### Node Details
+
+#### HubSpot — Lookalike Creator
+- **Type and technical role:** `n8n-nodes-base.hubspot`; intended CRM write-back node.
+- **Configuration choices:**  
+  - Authentication: HubSpot app token
+  - `email` parameter is currently set to `=`, which is incomplete
+  - No clear operation or property mapping is configured in the visible parameters
+- **Key expressions or variables used:**  
+  - Intended to use enriched email, but no valid expression is present
+- **Input and output connections:**  
+  - Input from **AI Agent (Lookalike)**
+  - No downstream nodes
+- **Version-specific requirements:** `typeVersion: 2.2`
+- **Edge cases or potential failure types:**  
+  - Incomplete configuration will likely fail at runtime
+  - AI output may not include the email or normalized profile fields needed for upsert unless merged beforehand
+  - Custom properties like `recommended_action` and `brand_fit` must exist in HubSpot before mapping
+- **Sub-workflow reference:** None.
+
+---
+
+## 3. Summary Table
+
+| Node Name | Node Type | Functional Role | Input Node(s) | Output Node(s) | Sticky Note |
+|---|---|---|---|---|---|
+| HubSpot Trigger | n8n-nodes-base.hubspotTrigger | Entry point on HubSpot contact creation |  | Get Contact by ID | ## HubSpot Trigger<br>Fires on `contact.creation`. Outputs `contactId` for the next node.<br>Add `contact.propertyChange` in HubSpot app settings to also catch updates. |
+| Get Contact by ID | n8n-nodes-base.hubspot | Fetch HubSpot contact fields from contactId | HubSpot Trigger | Extract Email | ## Get Contact by ID<br>Fetches `firstname`, `company`, `hs_full_name_or_email` from HubSpot using `contactId`.<br>Add more properties to `propertiesCollection` if your CRM has custom creator fields. |
+| Extract Email | n8n-nodes-base.set | Map HubSpot data to simplified fields | Get Contact by ID | Enrich by Email | ## Extract Email<br>Maps HubSpot fields into clean `Email` and `Company Name` variables. |
+| Enrich by Email | n8n-nodes-influencersclub.influencersClub | Enrich seed contact through influencers.club email lookup | Extract Email | Normalize; Stop and Error | ## Enrich by Email<br>POSTs email to influencers.club and returns a full main profile info.<br>Error → Stop and Error node. |
+| Stop and Error | n8n-nodes-base.stopAndError | Stop execution on failed seed enrichment | Enrich by Email |  | ## Stop and Error<br>Triggered when enrichment fails or returns no creator. Halts execution with a logged error message. |
+| Normalize | n8n-nodes-base.code | Normalize seed creator profile and detect main platform | Enrich by Email | Filter - Is Creator? | ## Normalize (Code in JS)<br>Detects `main_platform` by comparing follower counts across all platform blocks. Computes `follower_tier`, `engagement_tier`, `growth_trend`. Strips CloudFront URL signing params.<br>Outputs a flat clean object for the filter and lookalike nodes. |
+| Filter - Is Creator? | n8n-nodes-base.if | Keep only creator profiles | Normalize | Find Similar Creators | ## Filter — Is Creator?<br>Checks `is_creator === true`. Only real creators proceed.<br>**Port 0** → Lookalike API<br>**Port 1** → dead end, dropped silently |
+| Find Similar Creators | n8n-nodes-influencersclub.influencersClub | Search for similar creators | Filter - Is Creator? | Split Out | ## Lookalike API<br>Finds similar creators using `main_platform` + `handle`. Returns up to 5 accounts.<br>Adjust `paging.limit` (max ~50) or add filters: `engagement_percent.min`, `number_of_followers`, `has_done_brand_deals: true`. |
+| Split Out | n8n-nodes-base.splitOut | Split lookalike accounts array into individual items | Find Similar Creators | Loop Over Items (Lookalike) | ## Lookalike API<br>Finds similar creators using `main_platform` + `handle`. Returns up to 5 accounts.<br>Adjust `paging.limit` (max ~50) or add filters: `engagement_percent.min`, `number_of_followers`, `has_done_brand_deals: true`. |
+| Loop Over Items (Lookalike) | n8n-nodes-base.splitInBatches | Process lookalikes sequentially | Split Out; AI Agent (Lookalike) | Enrich by Handle (Full) | ## Loop Over Items<br>Processes one lookalike at a time to avoid API rate limits.<br>**Port 0** → all done, exits loop<br>**Port 1** → next item → Enrichment API |
+| Enrich by Handle (Full) | n8n-nodes-influencersclub.influencersClub | Fetch full lookalike creator profile | Loop Over Items (Lookalike) | Normalize Profile (Lookalike) | ## Enrich Lookalike<br>Fetches full profile per lookalike: post data, audience, income, connected platforms.<br>Error output suppressed — failed enrichments skip silently so the loop continues. |
+| Normalize Profile (Lookalike) | n8n-nodes-base.code | Convert lookalike data into standardized nested schema | Enrich by Handle (Full) | AI Agent (Lookalike) | ## Enrich Lookalike<br>Fetches full profile per lookalike: post data, audience, income, connected platforms.<br>Error output suppressed — failed enrichments skip silently so the loop continues. |
+| AI Agent (Lookalike) | @n8n/n8n-nodes-langchain.agent | Produce structured creator evaluation | Normalize Profile (Lookalike); OpenAI Model (Lookalike)1; Structured Output Parser (Lookalike)1 | HubSpot — Lookalike Creator; Loop Over Items (Lookalike) | ## AI Agent<br>GPT-4o-mini analyzes the profile and returns structured JSON: summary, strengths, concerns, audience authenticity, growth trajectory, niche clarity, brand fit, recommended action, and estimated reach per post.<br>Edit the system prompt to add campaign context or scoring bias. |
+| OpenAI Model (Lookalike)1 | @n8n/n8n-nodes-langchain.lmChatOpenAi | Provide GPT-4o-mini model to agent |  | AI Agent (Lookalike) | ## AI Agent<br>GPT-4o-mini analyzes the profile and returns structured JSON: summary, strengths, concerns, audience authenticity, growth trajectory, niche clarity, brand fit, recommended action, and estimated reach per post.<br>Edit the system prompt to add campaign context or scoring bias. |
+| Structured Output Parser (Lookalike)1 | @n8n/n8n-nodes-langchain.outputParserStructured | Enforce JSON schema on AI output |  | AI Agent (Lookalike) | ## AI Agent<br>GPT-4o-mini analyzes the profile and returns structured JSON: summary, strengths, concerns, audience authenticity, growth trajectory, niche clarity, brand fit, recommended action, and estimated reach per post.<br>Edit the system prompt to add campaign context or scoring bias. |
+| HubSpot — Lookalike Creator | n8n-nodes-base.hubspot | Create or update lookalike in HubSpot | AI Agent (Lookalike) |  | ## HubSpot Upsert<br>Creates or updates the lookalike as a HubSpot contact using their enriched email.<br>Map `recommended_action`, `brand_fit`, and `estimated_reach_per_post` to custom HubSpot properties for outreach sequencing. |
+| 📋 OVERVIEW | n8n-nodes-base.stickyNote | Workflow documentation note |  |  | # HubSpot → Lookalike Creator Discovery<br>**How it works** When a contact is created in HubSpot, this pipeline checks if they're a creator, finds similar creators via the influencers.club API, enriches and AI-analyzes each one, then pushes qualified lookalikes back into HubSpot.<br>**Flow:** Trigger → Fetch contact → Extract email → Enrich → Detect platform → Filter creators → Lookalike API → Loop each result → Enrich → Normalize → AI analysis → HubSpot upsert<br>**Platform detection:** `main_platform` is auto-detected by comparing follower counts across TikTok / Instagram / YouTube / Twitter — the highest audience wins. Non-creators are silently dropped at the Filter node.<br>**Set up required**<br>- HubSpot Developer App (trigger)<br>- HubSpot App Token (read + write)<br>- influencers.club API key<br>- OpenAI API key |
+| Note: Trigger | n8n-nodes-base.stickyNote | Documentation note |  |  | ## HubSpot Trigger<br>Fires on `contact.creation`. Outputs `contactId` for the next node.<br>Add `contact.propertyChange` in HubSpot app settings to also catch updates. |
+| Note: Get Contact | n8n-nodes-base.stickyNote | Documentation note |  |  | ## Get Contact by ID<br>Fetches `firstname`, `company`, `hs_full_name_or_email` from HubSpot using `contactId`.<br>Add more properties to `propertiesCollection` if your CRM has custom creator fields. |
+| Note: Extract Email | n8n-nodes-base.stickyNote | Documentation note |  |  | ## Extract Email<br>Maps HubSpot fields into clean `Email` and `Company Name` variables. |
+| Note: Enrich by Email | n8n-nodes-base.stickyNote | Documentation note |  |  | ## Enrich by Email<br>POSTs email to influencers.club and returns a full main profile info.<br>Error → Stop and Error node. |
+| Note: Stop and Error | n8n-nodes-base.stickyNote | Documentation note |  |  | ## Stop and Error<br>Triggered when enrichment fails or returns no creator. Halts execution with a logged error message. |
+| Note: Normalize Seed | n8n-nodes-base.stickyNote | Documentation note |  |  | ## Normalize (Code in JS)<br>Detects `main_platform` by comparing follower counts across all platform blocks. Computes `follower_tier`, `engagement_tier`, `growth_trend`. Strips CloudFront URL signing params.<br>Outputs a flat clean object for the filter and lookalike nodes. |
+| Note: Filter | n8n-nodes-base.stickyNote | Documentation note |  |  | ## Filter — Is Creator?<br>Checks `is_creator === true`. Only real creators proceed.<br>**Port 0** → Lookalike API<br>**Port 1** → dead end, dropped silently |
+| Note: Lookalike API | n8n-nodes-base.stickyNote | Documentation note |  |  | ## Lookalike API<br>Finds similar creators using `main_platform` + `handle`. Returns up to 5 accounts.<br>Adjust `paging.limit` (max ~50) or add filters: `engagement_percent.min`, `number_of_followers`, `has_done_brand_deals: true`. |
+| Note: Loop | n8n-nodes-base.stickyNote | Documentation note |  |  | ## Loop Over Items<br>Processes one lookalike at a time to avoid API rate limits.<br>**Port 0** → all done, exits loop<br>**Port 1** → next item → Enrichment API |
+| Note: Enrich Lookalike1 | n8n-nodes-base.stickyNote | Documentation note |  |  | ## Enrich Lookalike<br>Fetches full profile per lookalike: post data, audience, income, connected platforms.<br>Error output suppressed — failed enrichments skip silently so the loop continues. |
+| Note: AI Agent | n8n-nodes-base.stickyNote | Documentation note |  |  | ## AI Agent<br>GPT-4o-mini analyzes the profile and returns structured JSON: summary, strengths, concerns, audience authenticity, growth trajectory, niche clarity, brand fit, recommended action, and estimated reach per post.<br>Edit the system prompt to add campaign context or scoring bias. |
+| Note: HubSpot Out | n8n-nodes-base.stickyNote | Documentation note |  |  | ## HubSpot Upsert<br>Creates or updates the lookalike as a HubSpot contact using their enriched email.<br>Map `recommended_action`, `brand_fit`, and `estimated_reach_per_post` to custom HubSpot properties for outreach sequencing. |
+| Sticky Note1 | n8n-nodes-base.stickyNote | Documentation note |  |  | ## Find similar creators to your top performers and boost your brand visibility<br>**Step by step workflow to discover lookalike creators with multi social (Instagram, Tiktok, Youtube, Twitter, Onlyfans, Twitch and more) data using the influencer.club API and add them to Hubspot**.<br>https://influencers.club/creatorbook/find-similar-creators-to-top-performers/ |
+
+---
+
+## 4. Reproducing the Workflow from Scratch
+
+1. **Create a new workflow** in n8n and name it something like:  
+   **Find and add similar creators to HubSpot based on top performers**.
+
+2. **Add a HubSpot Trigger node**.
+   - Node type: **HubSpot Trigger**
+   - Credentials: **HubSpot Developer API**
+   - Configure it for the contact creation event.
+   - Optionally add contact property change events if you also want updates.
+   - Set `maxConcurrentRequests` to `5`.
+
+3. **Add a HubSpot node named “Get Contact by ID”**.
+   - Node type: **HubSpot**
+   - Operation: **Get Contact**
+   - Authentication: **App Token**
+   - Credentials: HubSpot app token with contact read access
+   - Set Contact ID to:
+     `{{ $json.contactId }}`
+   - Request the needed properties. At minimum, include:
+     - `email`
+     - `firstname`
+     - `company`
+     - `hs_full_name_or_email`
+   - If you depend on associated company information, also ensure your retrieval approach actually returns associated company data, or add an additional HubSpot association/company node.
+
+4. **Connect HubSpot Trigger → Get Contact by ID**.
+
+5. **Add a Set node named “Extract Email”**.
+   - Create fields:
+     - `Email` = `{{ $json.properties.email.value }}`
+     - `Company Name` = if available from the payload, otherwise map from a valid company source
+   - Important: the provided workflow assumes `associated-company.properties.name.value`, which may not exist unless explicitly fetched elsewhere. If needed, replace with a safer expression.
+
+6. **Connect Get Contact by ID → Extract Email**.
+
+7. **Add an influencers.club node named “Enrich by Email”**.
+   - Node type: **influencersClub**
+   - Credentials: influencers.club API key
+   - Operation: enrich by email
+   - Set email to:
+     `{{ $json.Email }}`
+   - Enable **Continue On Fail / Error Output** behavior so failed enrichments can be routed.
+
+8. **Connect Extract Email → Enrich by Email**.
+
+9. **Add a Stop and Error node**.
+   - Node type: **Stop and Error**
+   - Error message:
+     `API enrichment failed - contact is not a creator or API error`
+
+10. **Connect the error output of Enrich by Email → Stop and Error**.
+
+11. **Add a Code node named “Normalize”**.
+   - Paste the seed normalization JavaScript from the workflow.
+   - This code:
+     - unwraps the enrichment response
+     - compares follower counts across TikTok, Instagram, YouTube, and Twitter
+     - sets `main_platform`
+     - computes formatted follower count, tiers, growth trend
+     - creates a flattened creator object
+
+12. **Connect Enrich by Email main output → Normalize**.
+
+13. **Add an IF node named “Filter - Is Creator?”**.
+   - Condition:
+     - left value: `{{ $json.is_creator }}`
+     - operation: boolean equals
+     - right value: `true`
+
+14. **Connect Normalize → Filter - Is Creator?**.
+
+15. **Leave the false output unconnected** if you want silent dropping of non-creators.
+
+16. **Add an influencers.club node named “Find Similar Creators”**.
+   - Resource: `creator`
+   - Operation: `findLookalikes`
+   - Platform:
+     `{{ $json.main_platform }}`
+   - Filter key: `username`
+   - Filter value:
+     `{{ $json.handle }}`
+   - Credentials: influencers.club API key
+   - Optionally add advanced filters such as:
+     - follower thresholds
+     - engagement minimums
+     - brand deal requirement
+     - paging limit
+
+17. **Connect the true branch of Filter - Is Creator? → Find Similar Creators**.
+
+18. **Add a Split Out node named “Split Out”**.
+   - Field to split: `accounts`
+
+19. **Connect Find Similar Creators → Split Out**.
+
+20. **Add a Split In Batches node named “Loop Over Items (Lookalike)”**.
+   - Use default options for sequential processing.
+
+21. **Connect Split Out → Loop Over Items (Lookalike)**.
+
+22. **Add another influencers.club node named “Enrich by Handle (Full)”**.
+   - Resource: `creator`
+   - Operation: `enrichByHandle`
+   - Handle:
+     `{{ $json.profile.username }}`
+   - Platform:
+     `{{ $('Normalize').item.json.main_platform }}`
+   - Enable continue-on-error behavior so one failed lookalike does not stop the whole run.
+
+23. **Connect output 1 of Loop Over Items (Lookalike) → Enrich by Handle (Full)**.
+   - In this pattern, output 1 is used for each batch item.
+
+24. **Add a Code node named “Normalize Profile (Lookalike)”**.
+   - Set mode to **Run Once for Each Item**.
+   - Paste the provided normalization script.
+   - This code standardizes multiple possible influencers.club response formats into `normalized_profile`.
+
+25. **Connect Enrich by Handle (Full) → Normalize Profile (Lookalike)**.
+
+26. **Add an OpenAI Chat Model node named “OpenAI Model (Lookalike)1”**.
+   - Node type: LangChain OpenAI Chat Model
+   - Credentials: OpenAI API key
+   - Model: `gpt-4o-mini`
+
+27. **Add a Structured Output Parser node named “Structured Output Parser (Lookalike)1”**.
+   - Use manual schema mode.
+   - Paste the full JSON schema from the workflow.
+   - Ensure all required fields and enums match exactly.
+
+28. **Add an AI Agent node named “AI Agent (Lookalike)”**.
+   - Prompt type: define manually
+   - Main prompt should inject:
+     `{{ JSON.stringify($json.normalized_profile, null, 2) }}`
+   - Add the system message instructing the model to return valid JSON only.
+   - Enable output parser usage.
+
+29. **Connect Normalize Profile (Lookalike) → AI Agent (Lookalike)**.
+
+30. **Connect OpenAI Model (Lookalike)1 to the AI Agent’s language model input**.
+
+31. **Connect Structured Output Parser (Lookalike)1 to the AI Agent’s output parser input**.
+
+32. **Add a HubSpot node named “HubSpot — Lookalike Creator”**.
+   - Authentication: HubSpot app token
+   - Configure it as create or update contact, depending on your CRM design.
+   - Recommended matching key: email, if enrichment returns one.
+   - Map fields such as:
+     - email
+     - firstname/full name
+     - social handle
+     - platform
+     - follower count
+     - recommended action
+     - summary
+     - brand fit
+     - estimated reach per post
+   - Create corresponding custom HubSpot properties beforehand.
+
+33. **Connect AI Agent (Lookalike) → HubSpot — Lookalike Creator**.
+
+34. **Connect AI Agent (Lookalike) back to Loop Over Items (Lookalike)**.
+   - This closes the loop so the next lookalike item is processed.
+
+35. **Add sticky notes** for documentation if desired, mirroring:
+   - workflow overview
+   - trigger notes
+   - seed enrichment notes
+   - creator filtering
+   - lookalike discovery
+   - sequential processing
+   - AI analysis
+   - HubSpot output mapping
+
+36. **Validate the workflow carefully before activation**.
+   - Confirm `email` is really available from HubSpot.
+   - Confirm `associated-company` references are valid or replaced.
+   - Confirm the final HubSpot node has a real email mapping instead of an empty expression.
+   - Confirm the AI output still contains the fields you need, or add a Merge node if you must preserve both normalized profile data and AI analysis together.
+
+37. **Recommended hardening adjustments before production**:
+   - Add a Merge node after AI analysis if you need both source profile data and AI assessment in HubSpot.
+   - Add deduplication before HubSpot create/update using `dedup_key`.
+   - Add an IF node to skip lookalikes with no email if HubSpot requires email-based upsert.
+   - Add retry logic or throttling for API calls if you expect volume.
+
+### Required credentials
+- **HubSpot Developer API** for the trigger
+- **HubSpot App Token** for contact get/create/update
+- **influencers.club API credential**
+- **OpenAI API credential**
+
+### Input/output expectations
+- **Trigger input:** HubSpot contact creation event
+- **Seed enrichment input:** email address from HubSpot contact
+- **Lookalike discovery input:** normalized `main_platform` and `handle`
+- **AI input:** `normalized_profile`
+- **HubSpot output input:** ideally merged enriched profile + AI analysis + valid email
+
+---
+
+## 5. General Notes & Resources
+
+| Note Content | Context or Link |
+|---|---|
+| Find similar creators to your top performers and boost your brand visibility | influencers.club article |
+| Step by step workflow to discover lookalike creators with multi social (Instagram, Tiktok, Youtube, Twitter, Onlyfans, Twitch and more) data using the influencer.club API and add them to Hubspot | https://influencers.club/creatorbook/find-similar-creators-to-top-performers/ |
+| Required setup includes HubSpot Developer App, HubSpot App Token, influencers.club API key, and OpenAI API key | Workflow overview note |
+| Platform detection is automatic based on the highest follower count among TikTok, Instagram, YouTube, and Twitter | Workflow overview note |
+| Non-creators are dropped silently at the filter step | Workflow overview note |
+
+### Important implementation note
+The final HubSpot node appears incomplete in the provided workflow JSON. To make the workflow operational, you will likely need to finish the HubSpot create/update configuration and ensure the node receives both a valid contact identifier (usually email) and the enriched/AI-analyzed fields you want to store.