The Querit Python SDK provides a convenient way to interact with the Querit Content Retrieval Platform. It offers:

• Simple search interface for content retrieval
• Type-annotated request/response models
• Error handling for API responses
• Support for various search parameters and filters

For the Querit Content Retrieval Platform, we provide a Python SDK (Querit SDK) that allows developers to easily integrate and use Querit's content search capabilities programmatically.

Installation

Requirements

• Python 3.7+
• pip package manager

Install from PyPI

pip install querit

Install from source

git clone https://github.com/querit-ai/querit-python.git
pip install -e .

Verify Installation

python3 -c "import querit; print(querit.__version__)"

Quick Start

Authentication

First, obtain your API key from the Querit platform.

Basic Usage

from querit import QueritClient
from querit.models.request import SearchRequest
from querit.errors import QueritError

# Initialize client
client = QueritClient(
    api_key="Bearer your_api_key_here",
    timeout=30  # Optional timeout in seconds
)

# Create search request
request = SearchRequest(
    query="chat",
    count=5,
    # Add more parameters as needed
)

try:
    # Execute search
    response = client.search(request)

    # Process results
    for item in response.results:
        print(f"Title: {item.title}")
        print(f"URL: {item.url}")
        print("-" * 50)

except QueritError as e:
    print(f"Search failed: {e}")

## Advanced Usage

### Customizing Search Requests
```python
from querit.models.request import SearchRequest

# Advanced search with filters
request = SearchRequest(
    query="machine learning",
    count=10,
    filters={
        "language": "english",
        "date_range": "d1"
    }
)

Error Handling

The SDK provides specific error classes:

• QueritAPIError: API request failures
• QueritAuthError: Authentication failures
• QueritValidationError: Invalid request parameters

Best Practices

1. Reuse client instances rather than creating new ones for each request
2. Set appropriate timeout values for your use case
3. Handle rate limiting by implementing retry logic
4. Cache frequently used search results when possible

AI-Powered Search Examples

These examples combine Querit with an OpenAI-compatible LLM to answer questions using live web search results. Both require the following environment variables:

export OPENAI_API_KEY="sk-..."
export OPENAI_BASE_URL="https://api.deepseek.com/v1"   # or any OpenAI-compatible endpoint
export OPENAI_MODEL="deepseek-chat"                    # or gpt-4o-mini, qwen-max, etc.
export QUERIT_API_KEY="qr-..."

AI Search with Summarization (ai_search_summary.py)

A three-step pipeline: the LLM first rephrases the question into optimized search keywords (XML output), then Querit executes the searches, and finally the LLM synthesizes a cited answer from the results.

python examples/ai_search_summary.py "GDP Data of Laos for the Past 10 Years"

AI Search via Tool Use (ai_search_tool_use.py)

Lets the LLM drive the search through function calling (tool use). In Round 1 the model decides which queries to run; in Round 2 it synthesizes a final answer from the retrieved results. No XML parsing needed — the model controls query decomposition autonomously and can issue multiple parallel searches in a single turn.

python examples/ai_search_tool_use.py "GDP Data of Laos for the Past 10 Years"

For more examples, see the examples/ directory in this repository.

Contact

If you experience any problems while using Querit, please feel free to reach out to us at support@querit.ai. Our team is ready to assist you.