Dashboard
Planner Qualifications Streams Generate RPL Pipeline Document Polish
Compliance Library Compliance Audit
Reference Chat
RTO Resource Library Book Builder Unit Clusters Units Upload Unit Templates Images
Settings Storage Pricing
Rules & Funding Downloads Learners Help

Image Generation

Configure AI-powered cover page images for your VET learning packs. Our image generator creates stunning, professional backgrounds that reflect your industry context.

🎨

★ How It Works

  • Images are generated using Google Imagen 4.0 Ultra
  • Prompts combine your master template with stream-specific scenes
  • Each unit gets a unique, contextual cover image
  • Fallback to Gemini 2.0 Flash if needed

📎 Connected to Runs

  • Cover images are generated in the Run Detail page
  • Use the Image Intensity slider when creating a run
  • Higher intensity = more detailed, professional images
  • Images are included in your ZIP download

🛠 Customisation

  • Edit the Master Prompt below to change style
  • Each Stream has its own scene description
  • Colour grades set the overall tone
  • Equipment keywords add industry-specific details

🌟 Best Practices

  • Keep prompts descriptive but concise
  • Use Australian industry terminology
  • Mention specific equipment for realism
  • Set intensity to 8-10 for best quality

Image Generation Workflow Visual Guide

1
New Run

Set intensity slider

➔
2
Processing

AI generates image

➔
3
Run Detail

Preview & regenerate

➔
4
Download

Image in ZIP

Tip: The image is generated during the content creation pipeline. Go to the Dashboard and click on any run to see the cover image section with preview and regeneration options.
Generate New Prompt: On the Run Detail page, use the "Generate New Prompt" button to create an AI-powered prompt tailored to the specific unit and stream. The AI uses your stream context (equipment, materials, processes) to generate accurate, industry-specific descriptions.

Master Prompt Template Global Settings

This template is used for all cover images. Placeholders are replaced with unit and stream-specific values.

Available Placeholders:
{stream_scene} {unit_code} {unit_title}

Loading master prompt template...

Stream Scene Templates Per-Stream

Each industry stream has its own scene description, colour grade, and equipment keywords that make images contextually relevant.

Note: Changes here affect all future image generations for that stream. Existing images won't change until you regenerate them.

Loading stream scenes...

Section Banner Generation 8-Banner System

Generate 16:9 section banners for your learning materials. Each stream gets 8 banners: 2 cycles (A/B variants) x 4 sections, providing visual variety.

Note: Banner generation may take 5-10 minutes per stream (8 banners). Each slot can retry up to 3 times if validation fails. Generated banners are saved as drafts and require approval before use.

House Style Guide & Prompt Templates

Download the guide to learn how to create custom banner prompts using ChatGPT. All prompts must follow the house style rules for consistent, high-quality images.

📄 Banner Prompt Guide

Complete guide with house style rules, section-specific requirements, ChatGPT prompt templates, and examples.

Download Guide (.docx)

📝 House Style Rules

  • No text, labels, logos, or signage
  • No identifiable faces (people from behind/blur)
  • 16:9 banner with negative space on left
  • Photorealistic industrial scenes
Creating Custom Prompts: Use the guide to ask ChatGPT to generate prompts for new streams or variations. Include the stream name, section type (orientation/process/WHS/troubleshooting), workplace context, and all house style rules.

🔄 How to Redo an Image or Run

1
Go to Dashboard

Find your completed run in the list

2
Open Run Detail

Click the run to see full details

3
Regenerate Image

Use the "Generate Cover" button with your custom prompt

4
Full Redo?

Create a new run from New Run page

Pro Tip: You can edit the prompt in the Run Detail page before regenerating to get exactly the image you want. Try adding specific details like lighting, camera angle, or atmosphere.

💬 Send Feedback

AI Assistant
Hi! I'm your AI assistant. Ask me anything about using this app, creating documents, or understanding the workflow.
Quick questions:
Compliance questions:
RTOFlow AI
Online General
Conversations

Dev Center V2

Drag to move
💡 Upload and manage instruction files for AI-assisted development. Drop files in the Inbox, then reference filenames in chat to trigger builds.
📄
Drop instruction files here or click to upload
Instruction Files
📭
No instruction files found
💡 Run context scans to understand the codebase at a glance. Each scan analyses a different aspect: system architecture, database tables, API endpoints, templates, and models.
Quick Context Scans

Loaded This Session

No context loaded yet

Session Activity

--:-- Session started. Run scans to populate context.

Scan Results

💡 Track development tasks through their lifecycle: Planned, In Progress, Implemented, Needs Testing, Tested, Verified, Released. Jobs require test evidence before verification.
Job Queue 0 items
📭
No jobs in queue
💡 Daily work log. Session activity is automatically logged here. Use "Enhance with AI" on entries to expand brief notes into detailed summaries.
Today's Work 0 entries
📝
No entries today
💡 View work history by date. Click "Close Out Today" to generate an AI-powered daily digest summarising what was accomplished.
Generate digest for today
Work History
📅
No history yet
💡 User feedback and AI-generated suggestions scored by priority. Convert high-priority items into tracked Jobs for the development pipeline.
Suggested Tasks
💭
Loading suggestions...
📥 User Requests
📭
Loading user requests...
💡 Reference guide for the Course Builder feature. Create custom training materials outside standard TGA units, with AI-powered conversational course design.
🎓 Course Builder FEATURE
🧠 Feature Overview
AI-powered course design and content generation pipeline for custom training materials outside TGA units. Organisations can design course outlines via AI conversation or manual entry, then generate full learner guides through the existing Stage C pipeline.
🔌 API Reference
Method Path Description
GET /api/course-builder/outlines List all outlines for org
POST /api/course-builder/outlines Create new outline
GET /api/course-builder/outlines/{id} Get outline with sections
PUT /api/course-builder/outlines/{id} Update outline metadata
DELETE /api/course-builder/outlines/{id} Delete draft outline
POST /api/course-builder/outlines/{id}/sections Add section
PUT /api/course-builder/sections/{id} Update section
DELETE /api/course-builder/sections/{id} Soft-delete section
POST /api/course-builder/outlines/{id}/reorder Reorder sections
POST /api/course-builder/outlines/{id}/generate Start generation pipeline
POST /api/course-builder/conversations Start AI design conversation
GET /api/course-builder/conversations List conversations
GET /api/course-builder/conversations/{id}/messages Get messages
POST /api/course-builder/conversations/{id}/messages Send message (triggers AI response)
POST /api/course-builder/conversations/{id}/approve Create outline from AI conversation
DELETE /api/course-builder/conversations/{id} Delete conversation
POST /api/course-builder/outlines/{id}/tga-framework Generate TGA framework for outline
GET /api/course-builder/outlines/{id}/tga-framework Get TGA framework data
GET /api/course-builder/outlines/{id}/tga-framework/export Export TGA framework as DOCX
POST /api/course-builder/outlines/{id}/tga-proposal Generate TGA proposal from framework
PATCH /api/course-builder/outlines/{id}/branding Update branding settings
POST /api/course-builder/outlines/{id}/tga-research Trigger TGA-specific research
GET /api/course-builder/outlines/{id}/tga-research Get TGA research results
PATCH /api/course-builder/outlines/{id}/tga-research Update TGA research decisions
⚙️ Pipeline Configuration
Feature Flag
course_builder — must be enabled per organisation via feature flags
Generation Pipeline
Course Builder runs skip Stage A/B. The blueprint is pre-built by the adapter, then goes straight to Stage C content generation.
AI Routing Stages
course_design — conversation AI for outline design
course_design_research — TGA-specific research framework for industry/regulatory context
Cost Tracking
All AI calls use cost_category="setup" and track via the unified AiUsageEvent system.
Adapter
app/services/course_builder_adapter.py converts CourseOutline → UnitBlueprint format for Stage C compatibility.
🗃️ Data Model
Table Description
course_outlines Main outline with metadata, status (draft → ready → generating → completed), AQF level, audience context, quality mode, tga_proposal_enabled, and rpl_pack_enabled flags
course_outline_sections Sections with learning outcomes, key topics, content type (theory/practical/assessment), complexity hint, and word targets
course_design_conversations AI design chat sessions linked to org/user, with context snapshot and status tracking
course_design_messages Individual messages (user/assistant/system) with interactive elements, user selections, and token usage tracking
💡 Agent skill registry. Skills are reusable instructions that extend what the AI agent can do. Auto-scans from the .agents/skills/ directory.
Agent Skills Registry
0
Installed
0
Planned
🧠
Loading skills...
💡 Project health dashboard: codebase metrics, pipeline success rates, AI spend tracking, and security status. Click Refresh All to update.
Project Statistics
📊
Loading stats...
RTOFlow Development Roadmap
AI-powered training and compliance platform for Australian RTOs. Tracking progress from concept through to production scale.
—%
—
Complete
—
Total Stages
—
Lines of Code
—
Files
—
AI Providers
—
Version
💡 SEO Master Plan tracker. Shows progress across all marketing and SEO initiatives, content coverage, and next priorities.
SEO Master Plan
📣
Loading marketing data...
💡 Quick-access shell commands for common operations. Click the copy button to copy a command, then paste it into the Replit Shell tab.
Shell Commands
🚀
Push to GitHub
Staged push — auto-batches large pushes