n8n-nodes-hixy-intelligence-cloud
v0.2.5
Published
An n8n node for the Hixy Intelligence Cloud. Hixy's advanced system is now accessible through the N8N node. It generates actionable data for decisions, ask a business question in plain language and get grounded recommendations, audience sentiments, conten
Maintainers
Readme
n8n-nodes-hixy-intelligence-cloud
An n8n node for the Hixy Intelligence Cloud.
Hixy's advanced system is now accessible through the N8N node. It generates actionable data for decisions: ask a business question in plain language and get grounded recommendations, audience sentiments, content ideas based on demand, opportunities, forecasts and trend intelligence — with confidence, a reasoning trace and safe source attribution. Built as a native, programmatic node with output modes, a dynamic industry picker, credit safety and idempotency. It's also usable as an AI Agent tool.
Install · Credentials · Resources · Output modes · Credit controls · Templates
Installation
Community-nodes install (recommended): Settings → Community Nodes → Install →
n8n-nodes-hixy-intelligence-cloud.
Local development:
npm install --legacy-peer-deps # starter toolchain peer-dep workaround
npm run build
npm run lint
npm run dev # local n8n with the node linkedCredentials
Create an API key in the developer console at web.hixy.app/hic → API Keys
(keys look like hxk_…). The Hixy Intelligence Cloud API credential:
| Field | Notes |
| --- | --- |
| API Key | Your hxk_… key. Sent as Authorization: Bearer …. |
| Environment | Production (default) or Custom. |
| Base URL | Only shown for Custom — a self-hosted / dedicated HIC deployment. |
The connection Test calls the authenticated GET /usage, so an invalid key fails here
(unlike /status, which is public).
Resources & operations
A single node with four resources, each mapped to a clear action (HIX = credits):
| Resource → Operation | Endpoint | Cost |
| --- | --- | --- |
| Intelligence → Ask HIC | POST /analyze | 1–10 HIX |
| Forecast → Forecast Trend or Industry | POST /forecast | 5 HIX |
| Trend → Get Trend Intelligence | GET /trend/{id} | 1 HIX |
| System → Get HIC Status | GET /status | Free |
Ask HIC
Provide a Question and an Industry (dynamic dropdown from the live catalogue, or a name/slug/expression). The engine classifies the question and returns an answer, ranked recommendations, confidence, reasoning, evidence and sources. Intent is auto-detected by default; override it under Options.
Forecast
Pick a Target (Trend / Topic / Industry), a Time Horizon (7d, 14d, 30d, 60d or 90d), and get
forward-looking movement. Simplify Output flattens it to current_stage / predicted_stage /
probability / confidence / risk / predicted_peak_date / predicted_lifespan — ideal for an IF node.
Get Trend
Full intelligence for one public trend by ID. Trend ID accepts an expression, so it chains
from Ask HIC (map its trend_id). Include-Sources / Include-Rationale prune client-side.
Get HIC Status
Liveness, data freshness and capabilities. Fail if Data Is Stale (+ Maximum Data Age) turns it into a guard that stops a workflow when HTE data is too old.
Output modes
Ask HIC emits results three ways:
- Full Response — one item = the whole analysis (storage/reports).
- One Item Per Recommendation — split each recommendation into its own item (best for IF / Slack / Notion automation).
- Top Recommendation Only — just the highest-ranked one.
Simplify Output returns clean, decision-friendly fields. Minimum Confidence and Maximum Items filter/cap client-side.
Credit controls
Because HIC spends HIX, the node makes cost visible and safe:
- Maximum HIX to Spend — a preflight (via
/analyze/estimate) that blocks a request before any HIX is spent if the estimate exceeds your limit. - Idempotency Key — a repeat with the same key replays the stored answer without re-charging. Auto-generated per item when blank, so retries never double-charge.
- On Insufficient Credits —
Stop Workflow(default) /Continue With Error Item/Skip Item. - Return Usage Metadata — attach
hix_used/hix_remainingto each item. - Cache Identical Requests — dedupe identical requests within one execution.
- Execute Once — process only the first input item (avoid spending HIX per incoming item).
Errors are mapped to actionable messages: 401 invalid key, 402 shows required vs available
HIX, 403 scope, 404 not found, 429 respects Retry-After (auto-retried with backoff+jitter;
free GET on 429/5xx, POST only on 429/502/503/504, and the charged GET /trend on 429 only so a
post-charge failure is never re-charged).
Using it as an AI tool
The node sets usableAsTool: true. Under an AI Agent node the agent can call Ask HIC /
Forecast / Get Trend directly — HIC is AI-native (returns a synthesized, grounded answer).
Templates
Importable starter workflows live in templates/ — in n8n use Import from File:
01-weekly-trend-alert— Schedule → Ask HIC (Trends) → IF confidence ≥ 0.8 → alert02-opportunity-monitor— Schedule → Ask HIC (Opportunities) → IF opportunity ≥ 75 → store03-weekly-content-plan— Schedule → Ask HIC (Content Ideas) → draft04-forecast-watch— Schedule → Forecast → IF probability > 0.75 → alert05-trend-enrichment— Ask HIC → Get Trend (chained) → store06-multi-industry-report— Industry list → loop → Ask HIC → aggregate
Compatibility
- n8n with community-node support,
n8nNodesApiVersion1. - Built + linted with
@n8n/node-cli(strict). No runtime deps beyond peern8n-workflow.
Resources
- Developer console & docs: https://web.hixy.app/_hte/console/docs
- n8n community nodes: https://docs.n8n.io/integrations/#community-nodes
