> ## Documentation Index
> Fetch the complete documentation index at: https://docs.hiveku.com/llms.txt
> Use this file to discover all available pages before exploring further.
# Workflow Node Reference
> Every node category in the workflow editor, with highlights and limits
The workflow editor ships more than 350 node types. Click **Add Node** on the canvas (or drag from the palette) to browse them — the modal groups every node into the categories below, with triggers first, and you can search by name or description from the top of the modal.
This page is a map of the categories, not a field-by-field manual. Every node shows its own one-line description in the picker and a configuration panel with inline hints once placed.
## Triggers
Every workflow starts with exactly one trigger. The families:
| Family | Examples |
| --------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Core** | Inbound Webhook, Scheduled Trigger (cron), Manual Trigger, Website Visitor |
| **Forms** | Form Submitted — fires for every form on your sites, whether embedded, hosted, or captured from a deployed page. Spam and duplicate submissions are filtered out before the trigger fires |
| **SMS** | SMS Received — fires on inbound texts, with optional keyword and number filters |
| **CRM** | Contact created/updated, deal stage changes, and other pipeline events |
| **Bookings** | Booking Created, Booking Cancelled, Booking Rescheduled |
| **Helpdesk** | Ticket Created, Ticket Updated, Ticket Assigned, Ticket Resolved, New Message |
| **Billing** | Estimate, Invoice, Payment, Subscription, and Signature triggers — each filterable by event (sent, viewed, paid, overdue, refunded, and so on) |
| **Surveys** | Survey Response Received — filter by survey, score bucket (promoter/detractor), or score range |
| **Email & marketing** | Inbound email, campaign and sequence events |
| **Database** | Row inserted, updated, or deleted in a project database |
| **Shopify** | Orders, reviews, and subscription lifecycle events (including Billing Failed) |
The **SMS Received** trigger can be scoped with **To this number** so it only fires for texts sent to one of your numbers, plus an optional keyword filter. The trigger payload carries the full message text, the matched CRM contact, and any MMS attachments via `{{trigger.media}}`.
## Action categories
| Category | What's in it |
| -------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **AI & Agents** | AI Agent nodes — route to a department agent (with your brand context and tools) or a bare model for quick classification. Supports structured JSON output and sub-agent delegates |
| **CRM** | Create/update/find contacts, companies, deals, activities, and reports |
| **Helpdesk** | Open tickets, reply, assign, set status and priority |
| **Project Management** | Tasks, subtasks, status updates |
| **SEO** | Rankings, backlinks, content gaps, audits |
| **PPC & Ads** | Campaign metrics, pause/resume, budget changes — writes are staged for approval by default |
| **Social Media** | Create and schedule posts, approvals, analytics |
| **Marketing & Outbound** | Content, knowledge base, campaigns, sequences, audiences |
| **Analytics** | Pageviews, sessions, visitor journeys |
| **Builder & Deploy** | Projects, branches, commits, deploys, preview PRs |
| **Email** | Send, reply, search, and label across connected Gmail and Outlook accounts |
| **Calendar & Bookings** | Events, bookings, free-slot search |
| **CMS** | Read, write, search, promote, and bulk-import content entries (see below) |
| **Notifications** | Slack, Discord, Teams, push, in-app |
| **Database** | CRUD, bulk operations, SQL, and lookups on project databases |
| **HTTP** | API calls with retries, circuit breakers, batching, and rate limits |
| **Cache & State** | Cross-node caching, shared state, counters |
| **Files & Storage** | CSV/JSON parsing, uploads, downloads, URL fetch |
| **Data Transform** | Transform, validate, format, array and string utilities |
| **Flow Control** | Condition (If), Switch, For Each, Delay, subworkflows, wait/approval steps |
| **Mission Control** | Task lanes, intake triage, transitions, SLA breaches |
| **Hiveboards** | Visual boards, elements, sitemap scaffolds |
| **Creative Studio** | AI image and video generation, stock photos, media library, design exports (see below) |
| **Web Research** | Search, scrape, crawl, extract, and map the public web (see below) |
| **Voice** | Calls, recordings, numbers, extension health (see below) |
| **Accounting & Billing** | Vendors, bills, payments, time entries, AR/AP aging, P\&L (see below) |
| **Health & Safety Checks** | Integration health, deploy preflight, checkpoints, tracking diagnosis (see below) |
## Newest categories
### Creative Studio
Generate media inside a workflow and pass the result straight to a CMS or social step:
* **Generate Image** — one AI image per run; returns a public URL and a Media Library asset. Costs one image credit per run.
* **Generate Image Set** — up to 10 images in one call (hero, avatar, before/after variants). Bills per image.
* **Generate Video** — one short AI clip per run, with a monthly spending cap. Use **For Each** to produce several.
* **Stock Photo Search** — free, read-only search across Unsplash and Pexels; returns image URLs plus attribution text.
Generation nodes spend credits every time they run. Pair them with a Condition so a retriggered workflow doesn't re-render media it already produced, and check your run history after enabling.
### Web Research
* **Web: Search** — live web search with ranked results and page text, ready for a downstream AI node.
* **Web: Scrape Page** — one public page as clean markdown with title and links.
* **Web: Crawl Site** — a site section as markdown, capped at 50 pages per node.
* **Web: Extract Data** — structured fields from up to 10 public pages using a prompt or JSON schema.
* **Web: Map Site URLs** — fast URL discovery via sitemaps, without scraping. Feed the list into a loop.
Only public `http(s)` URLs are accepted; internal hosts are refused.
### CMS
The CMS family covers the full content lifecycle — see [CMS overview](/cms/overview) for the underlying concepts:
* **Write Entry** creates entries as drafts by default; **Promote Draft** publishes them. This lets an AI-written article wait for a QA step before going live.
* **Read Entry**, **Search Entries**, **List Entries**, and **List Collections** let workflows branch on existing content (for example, a dedupe check before writing).
* **Attach Image** adds an image to an entry field — typically `{{image-node.imageUrls[0]}}` from a Generate Image step.
* **Delete Entry** is a soft delete; tick **Draft** to discard only the draft and leave the published entry alone.
* **Bulk Import Entries** creates up to 200 entries from a JSON array; you choose skip-or-overwrite conflict behavior explicitly.
### Voice
Read-only visibility into your phone system: **List Calls** (with direction and disposition filters), **Get Call Detail** (recording, transcript, AI summary, sentiment), **List Numbers**, and **Extension Status**. None of these place calls or expose credentials — pair them with the **SMS** or notification nodes to act on what they find.
### Accounting & Billing
Action nodes that pair with the billing triggers: record bill and invoice payments, manage vendors and time entries, pull AR/AP aging and a cash-basis P\&L summary. Payment-recording nodes validate whole-cent amounts and support idempotency keys so a replayed run cannot book the same payment twice.
### Health & Safety Checks
Nodes for building self-monitoring automations:
* **Test Integration** — re-checks a connected integration's auth so a Condition can route dead connections to an alert.
* **Deploy Doctor** — read-only diagnosis of a deployed tier that isn't serving correctly.
* **Create Checkpoint** — snapshots every project file before a risky step; put it first in any workflow that edits a site.
* **Diagnose Tracking** — finds out why conversion tracking isn't recording.
* **Account Health Audit** — a drift-and-neglect snapshot with a score you can branch on.
## Side effects and Test Run
Nodes that send, publish, charge, or write externally are marked side-effecting. When you click **Test Run** in the editor, those nodes are simulated — the run inspector shows a **Simulated** badge with what the node *would have* done, and nothing is sent or billed. Real runs of the same nodes are protected by a duplicate-send guard, so a replayed run skips sends the original already completed.
## What's Next?
Triggers, merge variables, run history, and safeguards
Step-by-step editor walkthrough