Bring Your Own Model: Use Any AI Provider with VULK

Why One AI Model Isn't Enough

Different AI models excel at different things. Claude is exceptional at reasoning and complex architectures. GPT-4o is fast and reliable. DeepSeek offers incredible value. Groq delivers blazing inference speed. Mistral handles European languages beautifully.

So why would you lock yourself into just one?

With VULK's BYOM (Bring Your Own Model) system, you don't have to choose. On Pro, Team, Max, and Business, connect your own API keys and use whatever model fits your project.

Supported Providers

VULK supports 8 AI providers natively, plus any OpenAI-compatible endpoint:

Provider	Key Models	Best For
OpenAI	GPT-4o, GPT-4o-mini, o3	General-purpose, fast iteration
DeepSeek	DeepSeek V3, DeepSeek R1	Cost-effective, strong reasoning
Groq	Llama 3.3 70B, Mixtral	Ultra-fast inference (500+ tok/s)
Mistral	Mistral Large, Codestral	European languages, code generation
Together AI	Llama 3.1 405B, Qwen 2.5	Large open-source models
Fireworks	FireFunction, Llama 3.3	Production-grade, low latency
Ollama	Any local model	Privacy-first, offline development
Custom	Any OpenAI-compatible API	Self-hosted, enterprise setups

How It Works

Setting up BYOM takes about 30 seconds:

1. Go to Settings → API Keys in your VULK dashboard
2. Select your provider from the dropdown
3. Paste your API key — it's validated instantly
4. Choose your model from the auto-populated list
5. Start building — all generations route through your key

Your API key is encrypted with AES-256-GCM before storage. We never log your requests or responses. The key is only decrypted at the moment of making the API call.

Zero Credits, Full Power

When you use BYOM, model inference goes through your provider key. VULK doesn't charge generation credits for BYOM calls — you pay the provider's token costs, while VULK still provides orchestration, preview, validation, deployment, storage, and support.

This means:

• No credit limits on generation length or complexity
• No VULK credit burn for BYOM model calls
• Full model selection — use any model your provider offers
• Same VULK features — preview, deployment, file management all work the same

The Smart Routing Architecture

VULK's model router is intelligent about how it uses your key:

User Prompt → Intent Analysis → Model Router
                                    ├── Config files → Fast model (saves cost)
                                    ├── UI components → Medium model (balanced)
                                    └── Auth/Security → Best model (accuracy)

Even with BYOM, VULK's smart routing can optimize your costs by using lighter models for simpler files — if you enable it in settings.

Key Validation

Before saving your key, VULK makes a lightweight validation request to confirm:

• The key is valid and active
• The key has sufficient permissions
• The selected model is accessible

If validation fails, you get a clear error message explaining what to fix — no guessing.

Who Is BYOM For?

• Power users who generate dozens of apps per day and want predictable costs
• Enterprise teams with existing AI contracts and volume discounts
• Privacy-conscious developers who want to use local models via Ollama
• Open-source enthusiasts who prefer running Llama, Qwen, or Mixtral
• Cost optimizers who found a provider with better pricing for their usage

Getting Started

BYOM starts on Pro and is also available on Team, Max, and Business. Free and Builder use VULK-hosted models only. Head to Settings → API Keys, connect your provider, and start building with the model stack you already trust.

Your keys. Your models. Your apps. Available on production plans.