SUTRA with LiteLLM

This guide shows you how to use SUTRA models (V2 or R0) via LiteLLM, a powerful open-source library that lets you call different LLM providers with a unified interface.

📦 Step 1: Install Dependencies

pip install litellm

🔐 Step 2: Set Up API Key

Export your SUTRA API key as an environment variable:

export OPENAI_API_KEY="your_sutra_api_key"

Note: LiteLLM uses OPENAI_API_KEY for compatibility with OpenAI-style models.

⚙️ Step 3: Configure LiteLLM for SUTRA

Set up your call using SUTRA’s model and endpoint:

import litellm

response = litellm.completion(
    model="openai/sutra-v2",  # Or "openai/sutra-r0" for reasoning
    messages=[{"role": "user", "content": "Translate this to French: 'Good morning, how are you?'" }],
    api_base="https://api.two.ai/v2"  # SUTRA base URL
)

print(response['choices'][0]['message']['content'])

🧠 Use SUTRA-R0 for Reasoning-Based Prompts

response = litellm.completion(
    model="openai/sutra-r0",
    messages=[{"role": "user", "content": "If A is greater than B and B is greater than C, who is greatest?"}],
    api_base="https://api.two.ai/v2"
)

print(response['choices'][0]['message']['content'])

📎 Tips

• Use openai/sutra-v2 for:

  • Multilingual translation
  • Summarization
  • General Q&A

• Use openai/sutra-r0 for:

  • Reasoning
  • Logic-based queries
  • Legal/technical interpretation

• No extra configuration is needed beyond the api_base and model values.

🔗 Resources

• LiteLLM GitHub
• SUTRA API
• SUTRA GitHub Repository

Use LiteLLM + SUTRA for flexible, multilingual, and reasoning-capable AI applications through a single OpenAI-compatible API.