BS Net Trading AI Radio DJ - Complete System

AI Radio DJ Control Panel

BS Net Trading AI Agent Services

Dashboard Overview

Live Shows

Scheduled Shows

Available Presenters

Scripts Generated

Quick Actions

✨

Generate Script

Create presenter voiceover

📅

Create Show

Schedule a new program

👥

Manage Presenters

View all AI characters

Your AI Presenters (15 Total)

Live Control Panel

SYSTEM ONLINE

📻 NOW ON AIR

00:00:00

⏭️ UP NEXT

📈 TODAY'S STATS

Shows Aired

Scripts Used

Active Listeners

Uptime

00:00:00

⚡ QUICK ACTIONS

📋 ACTIVITY LOG

AI Presenters (15 Agents)

Show Schedule

⚡ Batch Script Generation

Pre-generate all scripts for scheduled shows in advance. Select a date range and the system will create all track intros, outros, promos, and time checks for every scheduled presenter.

From Date

To Date

Scripts Per Show

📝 Script Builder

Select Presenter

Script Type

Duration (seconds)

Context

📄 Script Preview

📝

No script generated yet

⚖️ OFCOM Compliance

Voice Output Manager

Total Generated

Processing

Ready to Export

Failed

Filter by Status

Filter by Presenter

Filter by Type

Prompt Engine

⚡ DEMO MODE — Placeholder scripts active

📡 Layer 1 — Station DNA

GLOBAL

This block applies to every script generated on Impact Digital Radio. It defines the station's overall brand voice, philosophy, and non-negotiable style rules.

🎭 Layer 2 — Character Sheet

PER PRESENTER

Auto-assembled from each presenter's profile + DynamoDB backstory. Shows what the AI sees when generating for this presenter.

Select Presenter

                            Select a presenter to load their character sheet...
                        

📝 Layer 3 — Script Type Instructions

PER SCRIPT TYPE

Each script type (intro, outro, weather, promo) has specific instructions that tell the AI what to include and how to structure it.

Script Type

                            Select a script type to see its instructions...
                        

💬 Layer 4 — Context

DYNAMIC

Additional context injected at generation time: track metadata (artist/title/duration), user input from the Script Builder, or RadioDJ queue data.

Example: Track Introduction Context

NEXT TRACK: "Yellow" by Coldplay (4:39)

Generate a track introduction for this specific song.
Reference the artist and title naturally.
Build anticipation without spoiling the experience.
Keep it under 30 seconds of spoken word.

Example: User Context from Script Builder

"Mention it's raining outside"

🔗 Composed Prompt Preview

What the AI actually receives →

🔗Select a presenter and script type in the Script Builder to see the full composed prompt here.

📦 Demo Script Output Log

⚡ Placeholder responses — see code comments for real API call

Demo outputs will appear here as scripts are generated

System Settings

Checking services...

🛡️ Backend Proxy

🧠 AI Script Engine

🎙️ ElevenLabs TTS

☁️ Azure TTS

🗄️ DynamoDB

🗺️ Voice ID Mapping

📻 RadioDJ Integration

🛡️ Backend Proxy

All secrets — API keys, AWS credentials, subscription tokens — live exclusively on your AWS backend. This frontend never sees or transmits them. Every service call routes through your Node.js proxy.

🔒

Architecture: Your proxy server (EC2 / Lambda) fetches secrets at runtime from AWS Secrets Manager or SSM Parameter Store. This panel stores only the proxy's public URL and a short-lived frontend auth token — nothing sensitive ever leaves the backend.

🌐 Proxy Base URL

Not connected

Endpoint

🔑 Frontend Auth Token

JWT · short-lived · auto-refresh

Issued by your backend at login to prove this session is authorised. The proxy validates it on every request.

Current Token

📋 Proxy Endpoints

Routes your backend exposes. Credentials are injected server-side before each upstream call.

POST/proxy/ai/generateClaude script gen

POST/proxy/tts/elevenlabsElevenLabs render

POST/proxy/tts/azureAzure TTS render

GET/proxy/db/character/:idSingle character

GET/proxy/db/charactersFull sync

GET/proxy/healthLiveness check

🧠 AI Script Engine

Configure which AI provider and model to use for script generation. API keys never leave your backend. Requests route through /proxy/ai/generate.

🔒

Key locations: AWS Secrets Manager → bsnet/anthropic-api-key and bsnet/openai-api-key. The proxy resolves the correct key based on your selected model and injects it into the upstream API call. Never cached on this client.

🤖 Model & Budget

Demo mode

AI Provider & Model

Max Tokens

Generation ModeLive API calls selected provider via proxy

🎙️ ElevenLabs TTS

Handles voice rendering for 11 presenters. API key is in Secrets Manager — only quality settings and test controls live here.

🔒

Key location: AWS Secrets Manager → bsnet/elevenlabs-api-key. Voice IDs are configured on the Voice ID Mapping tab and sent to the proxy alongside the script text at render time.

🎛️ Quality & Format

Not tested

Synthesis Model

Output Format

Test Presenter

☁️ Azure Cognitive TTS

Handles voice rendering for 5 presenters (Abby, Jody, Angie, Cindy, Tony). Subscription key and token refresh are managed entirely on the backend.

🔒

Key location: AWS Secrets Manager → bsnet/azure-tts-key. The proxy also handles the OAuth token refresh cycle with Azure's token endpoint automatically — no tokens hit this client.

🎛️ Region & Format

Not tested

Service Region

Output Format

Test Presenter

🗄️ DynamoDB

Stores backstories, speaking styles, and full character records. AWS credentials stay on the backend — only table names, region, and sync controls are configured here.

🔒

Credentials location: The EC2 instance role or Lambda execution role grants DynamoDB access via IAM — no access key ID or secret is stored anywhere in application code. The proxy uses the instance's ambient credentials automatically.

📦 Table & Region

Not tested

Characters Table Name

AWS Region

Test Character ID

🗺️ Voice ID Mapping

Maps each presenter to their voice ID on ElevenLabs or Azure. These IDs are sent to the proxy at render time — the proxy pairs them with the correct provider credentials server-side.

🎙️ ElevenLabs — 11 Presenters

☁️ Azure TTS — 5 Presenters

📻 RadioDJ Integration

Connect to your RadioDJ MySQL database to monitor the playlist queue and automatically inject AI voice tracks before songs play. Voice tracks are generated based on actual track metadata pulled from the queue.

🔒 How RadioDJ Integration Works

Your proxy server connects to the RadioDJ MySQL database and polls the queuelist table every few seconds. When it detects an upcoming track, it generates a personalized voice intro using the track's artist/title metadata, renders it with your presenter's voice, saves the MP3 to RadioDJ's media folder, and inserts it into the queue. All database credentials are stored in AWS Secrets Manager — the frontend never sees them.

🗄️ Database Connection

disconnected

Database Host

Port

Database Name

Username

Password

📋 Queue Monitor

Connect to database first to see upcoming tracks

⚙️ Automation Settings

stopped

Auto-Inject Voice Tracks

Lookahead Time (seconds)

Generate voice track this many seconds before the target song plays

Polling Interval (seconds)

How often to check the RadioDJ queue for changes

Media Folder Path

Where generated MP3s are saved before insertion into queue

📊 Integration Activity

Integration activity will appear here

Analytics

📝 Scripts Generated

+0 vs last period

🎙️ Voice Renders

+0 vs last period

🧠 AI Tokens Used

Est. cost: £0.00

⏱️ Total Audio

0:00

across all presenters

📊 Scripts Over Time

🏆 Top Presenters

🎬 Script Type Breakdown

💰 Token Budget

🎛️ Voice Provider Split

🕐 Recent Activity

AI Radio DJ Control Panel

Dashboard Overview

Quick Actions

Your AI Presenters (15 Total)

Live Control Panel

📻 NOW ON AIR

⏭️ UP NEXT

📈 TODAY'S STATS

⚡ QUICK ACTIONS

📋 ACTIVITY LOG

AI Presenters (15 Agents)

Show Schedule

⚡ Batch Script Generation

📝 Script Builder

📄 Script Preview

⚖️ OFCOM Compliance

Voice Output Manager

Prompt Engine

📡 Layer 1 — Station DNA

🎭 Layer 2 — Character Sheet

📝 Layer 3 — Script Type Instructions

💬 Layer 4 — Context

🔗 Composed Prompt Preview

📦 Demo Script Output Log

System Settings

🛡️ Backend Proxy

🧠 AI Script Engine

🎙️ ElevenLabs TTS

☁️ Azure Cognitive TTS

🗄️ DynamoDB

🗺️ Voice ID Mapping

📻 RadioDJ Integration

Analytics

Create New Show

Presenter Profile

Login

Shopping cart