GitHub Docs

Provider: SambaNova

SambaNova Provider

SambaNova runs AI models on custom RDU (Reconfigurable Dataflow Unit) hardware, delivering fast inference for open-source models.

Get an API Key

  1. Go to cloud.sambanova.ai
  2. Create an account or sign in
  3. Navigate to API Keys
  4. Create a new key and copy it

SambaNova offers a free tier with limited usage.

Configure VibeCody

Option 1: Environment variable (recommended)

export SAMBANOVA_API_KEY="..."
vibecli --provider sambanova

Option 2: Config file (~/.vibecli/config.toml)

[sambanova]
enabled = true
api_key = "..."
model = "Meta-Llama-3.1-70B-Instruct"

Model Selection

Model Strengths Best for
Meta-Llama-3.1-70B-Instruct Strong general coding Daily tasks (default)
Meta-Llama-3.1-8B-Instruct Ultra-fast Quick completions

Default: Meta-Llama-3.1-70B-Instruct

Override from the CLI:

vibecli --provider sambanova --model Meta-Llama-3.1-8B-Instruct

Best For

  • Fast inference – custom RDU hardware optimized for AI workloads
  • Open-source models – access Llama models with high throughput
  • Free tier – test fast inference without upfront costs

Verify Connection

vibecli --provider sambanova -c "Write a Python async web scraper"

Troubleshooting

Rate limited

Error: 429 Too Many Requests
  • Free tier has usage limits
  • Wait and retry, or upgrade for higher limits

Model not available