• Advanced Intelligent Routing: Utilize the power of GPT-4, Claude, and other leading LLMs with our advanced scoring functionality. Pulze.ai’s unique algorithm rates responses based on relevance, accuracy, lucidity, completeness, succinctness, appropriateness, and usefulness.
  • Custom Labels for Easy Filtering: Add custom labels as headers for your model requests. These labels will appear in the logs, making it easier for you to filter and locate specific requests.
  • Intelligent Routing Based on Your Preferences: Choose your key settings based on speed, cost, and quality. We make routing decisions according to your preference, ensuring that all requests made with that key follow your set heuristic.
  • Transparent Response Conversions: We uphold a clean interface with accurate conversions. If you send a request through our chat completion endpoint but we select a non-chat completion model, we’ll handle the conversion for you.
  • Detailed Response Metrics: We enrich your response data with extensive metrics, including unique ID, usage, costs, model selected, latency, tokens used, temperature, and cost savings compared to benchmark models.
  • Advanced Logging and Observability: Pulze.ai acts as your outsourced logging provider for all requests made through our platform, ensuring streamlined large language model operations.
  • Cost Efficiency: Save costs by letting us guide you to the most cost-effective solutions. Our platform tracks your costs and compares them to the leading top model to show you how much you’re saving.
  • Comprehensive Evaluation: With Pulze.ai, you secure thorough and automated evaluation of large language models. Save time and refine your choice for specific tasks and applications.
  • Pulze Playground: Experiment with different settings, balancing speed, cost, and quality in our playground environment. Get a feel for the capabilities of our model routing layer or embed to your homepage easily.
  • API Access: Generate a “secret key” to access our API. Equip your customers with access to the most advanced large language models (LLMs) currently available.
  • Source from Leading AI Labs: Our models are obtained from industry-leading AI research labs such as AI21 Labs, Anthropic, Cohere, OpenAI, and GooseAI, each designed to operate at peak levels considering diverse factors.
  • Prompt Execution and Inference: Enhance prompt generation, adjust pipelines, authenticate LLM outputs, and safeguard against prompt injection attacks with Pulze.ai. All of these benefits come standard with our platform.
  • Open source models: Integrate with any open-source model and utilize in your application.

Setting up and Using Pulze.ai API with OpenAI’s Python SDK

Here’s how to setup and use Pulze.ai API in your applications using OpenAI’s Python SDK (1.12.0):

import openai
openai.api_key = "<$PULZE_API_KEY>"
openai.base_url = "https://api.pulze.ai/v1/" # with trailing slash!

text_response = openai.completions.create(
  model="pulze-v0",
  prompt="Say Hello World!",
)

chat_response = openai.chat.completions.create(
  model="pulze",
  messages=[{
    "role": "user",
    "content": "Say Hello World!"
  }],
)

print(text_response, chat_response)

Using Pulze.ai with LangChain

You can also integrate Pulze.ai with LangChain as follows:

from langchain.models import OpenAI

pulze_model = OpenAI(
  openai_api_key="<$PULZE_API_KEY>",
  openai_api_base="https://api.pulze.ai/v1/"  # with trailing slash!
  model="pulze",
  max_tokens=256,
  temperature=0.4,
)

pulze_model.predict("Say Hello World!")