smart_toy
VoxGen
Multi-Provider AI Workspace

The AI workspace built for
Every Intelligence.

Chat, create, generate images, analyse documents — all in one platform. Powered by Claude, OpenAI, Gemini, xAI and loads of open-source and locally hosted models including Llama and more. Enterprise-grade security out of the box.

7+AI Providers
6+Platform Modules
∞Open Source Models
AES-256Encryption
grid_viewPlatform Modules

Everything you need,
nothing you don't.

Six production-ready modules working in harmony — from real-time AI chat to image generation, document intelligence, and knowledge bases.

chat

AI Chatbot

Real-time SSE streaming conversations with persistent session history. Switch models mid-conversation, toggle providers, markdown-rendered responses.

boltSSE Streaming historySession History codeMarkdown
description

Document Intelligence

Upload PDF, DOCX, TXT, and CSV files for AI-powered analysis. Summarise, extract key points, identify action items, draft replies, or generate executive reports.

picture_as_pdfPDF table_chartCSV analytics6 Modes
edit_note

Content Studio

Template-driven content generation with variable injection and 8-tone control. Built-in templates for blogs, emails, social posts, product descriptions, and reports.

dashboard_customize6 Templates tune8 Tones
image

Image Studio

Generate images with Stability AI SDXL or DALL-E 3. Choose from 10 style presets and 7 aspect ratios. AI prompt enhancement rewrites your prompt for better results.

auto_awesomeSDXL palette10 Styles psychologyAI Boost
tune

Prompt Builder

Visual prompt template editor with variable schema builder. Supports text, textarea, dropdown, number, and toggle fields with live auto-detection of placeholders.

buildVisual Builder shareShare Prompts
library_books

Knowledge Bases

Group documents into named knowledge stores for RAG-ready context. Attach, detach, and manage documents with chunk tracking and vector search readiness.

hubRAG-Ready searchVector Search
hubAI Provider Network

Every leading model,
one unified API.

All AI calls route through a smart dispatcher that selects the correct API format — OpenAI-compatible or native Anthropic. Switch providers without changing a line of code.

Anthropic
Claude 3 & 4 models with native SSE streaming. Nuanced reasoning and analysis.
ChatStreaming
OpenAI
GPT-4o, GPT-4 Turbo, and DALL-E 3 for best-in-class text and image generation.
ChatDALL-E 3
Google Gemini
Gemini Pro and Flash via OpenAI-compatible endpoint. Fast and multimodal out of the box.
ChatMultimodal
computer
Ollama / Local
Full local inference via Ollama. Zero API costs, zero data sent externally. Any GGUF model.
Self-HostedAny Model
developer_modeOpen Source & Local AI

Your data. Your models.
Your infrastructure.

VoxGen connects to HuggingFace's massive open-source model library and runs local models via Ollama — giving you full privacy, zero API costs, and complete control over your AI stack.

psychology
Llama 3.2
Meta's flagship open-source model. Exceptional reasoning, coding, and multilingual tasks. Available in 1B, 3B, 8B, and 70B sizes.
OllamaHuggingFaceChat
rocket_launch
Mistral 7B
Punches far above its weight class. Best-in-class for its parameter count with excellent instruction following and code generation.
OllamaHuggingFaceCode
science
Qwen 2.5
Alibaba's multilingual powerhouse. Exceptional at Chinese, Japanese, Korean alongside English. Strong coding and math capabilities.
OllamaMultilingual
diamond
Phi-4 Mini
Microsoft's compact reasoning model. Frontier-level reasoning in a tiny footprint — ideal for edge deployment.
OllamaReasoning
eco
Gemma 3
Google's open model family. Strong instruction following with safe, helpful responses. Optimised for efficiency.
OllamaHuggingFace
code
DeepSeek Coder
Purpose-built for code generation, completion, and explanation. Trained on 2 trillion code tokens across 87 programming languages.
HuggingFace87 langs
translate
NLLB-200
Meta's No Language Left Behind model. Translates between 200 languages with state-of-the-art accuracy.
HuggingFace200 langs
add_circle
Add Any Model
Paste any HuggingFace model URL into the admin panel. Metadata auto-resolves — name, type, context window, and gated status.
HuggingFace URLAuto-detect
shieldEnterprise Security

Security isn't an add-on.
It's the foundation.

Every layer of VoxGen is hardened — from password hashing to file storage, API key encryption, and audit logging. Built for compliance from day one.

Argon2IDPassword Hashing
AES-256Key Encryption
30Audit Events
0msSQL Injection Risk
password
Argon2ID Password Hashing
Memory-hard, GPU-resistant hashing — the gold standard for credential security. Session fixation prevented via regeneration on every login.
CWE-916
encrypted
AES-256-GCM API Key Encryption
All third-party API keys encrypted with a 256-bit app key and random 96-bit IV. Plaintext never touches the database.
FIPS 197
verified_user
CSRF Enforcement on Every Route
32-byte random tokens validated on all POST requests — both body and X-CSRF-Token header. Every fetch() call sends the token automatically.
OWASP A05
manage_accounts
3-Tier RBAC with Self-Protection
Super Admin, Admin, and User roles with strict route middleware. Users cannot modify their own role or escalate privileges.
Least Privilege
dataset
PDO Prepared Statements + Audit Log
Zero string concatenation in SQL. 30+ audited action types — auth, AI events, document operations, and admin actions — all immutable.
SQL Hardened
rocket_launchSetup in Minutes

From zero to AI-powered
in 4 steps.

Run the 5-step wizard. Environment checks, DB setup, and admin account creation in minutes — shared hosting compatible.

1
Install
Run the 5-step wizard. Environment checks, DB setup, and admin account creation in minutes — shared hosting compatible.
2
Connect AI
Add API keys for cloud providers or point to your Ollama/HuggingFace endpoint. Live connection test confirms everything works.
3
Invite Team
Create user accounts with role-based access. Super Admin, Admin, and User tiers ensure everyone has exactly the right permissions.
4
Start Creating
Chat, generate content, analyse documents, create images, and build prompt libraries. Everything logs to an immutable audit trail.
PHP 8.1+ MariaDB 10.4+ Apache / Nginx Shared Hosting Ollama Support

Pricing

Starter
R499/mo
3 users · AI Chat, Docs, Content Studio, Prompts
Most Popular
Professional
R1299/mo
10 users · All modules + Image Studio + Knowledge Bases + RAG
Enterprise
Custom
Unlimited users · White label · On-premise · SLA

Bring your own API keys — you only pay for what you use. Run Ollama locally for zero AI costs.

smart_toyVoxGen

One platform.
Every AI model.

Self-hosted. Your data stays on your infrastructure. Bring your own API keys — cloud giants, open-source, or fully local. No lock-in, ever.

rocket_launchDeploy VoxGen paymentsView Pricing support_agentGet Support