Skip to main content

A simple, powerful and highly-configurable Answer Engine.

Project description

AuxKnow 🛸

Khoj Logo

GitHub License Python Version Contributions Welcome

Answer Engines, everywhere.

💡 Docs   •   🕸️ Web   •  

AuxKnow is an advanced, state-of-the-art, fully customizable Answer Engine library, built to streamline the integration of intelligent answering capabilities into your applications. With AuxKnow, you can deploy feature-rich Answer Engines with minimal effort. For detailed guidance, refer to the Usage Instructions section in the documentation.


Background

At The Hackers Playbook, we are committed to building cutting-edge AI infrastructure and platforms grown out of Bharat (India). AuxKnow represents a significant milestone in our mission to deliver robust, developer-friendly tools for crafting AI-driven solutions. We encourage you to explore, innovate, and share your insights to help us improve and evolve. We look forward to your participation in our mission, since we are largely community driven and believe in Open Source culture.

AuxKnow leverages the capabilities of Perplexity’s Sonar models to enable seamless integration of "Answer Engine" functionality. While Perplexity’s models provide the foundational technology, AuxKnow bridges the gap by delivering a highly configurable and user-friendly library to accelerate application development.


Key Features

  • Fully Customizable: Configure AuxKnow to align perfectly with your application’s requirements. Its flexibility ensures it can adapt to various use cases, from simple Q&A to complex contextual reasoning.

  • Dynamic Query Routing: Automatically route user queries to the most appropriate model, ensuring optimal performance and accuracy.

  • Intelligent Query Restructuring: AuxKnow restructures user queries to improve response quality, delivering more accurate and relevant answers.

  • Contextual Search: AuxKnow can either leverage predefined contexts or autonomously build its own context to ensure that answers remain grounded in relevant knowledge.

  • Session Management: Seamlessly manage user interactions, including access to session history, context updates, and session resets. Export session data for analytics and debugging.

  • Streaming Responses: Enable real-time, incremental response generation for an enhanced user experience.

  • Source Attribution: Responses include detailed citations, ensuring transparency and reliability by grounding answers in trustworthy sources.


Installation

Install AuxKnow with a single command:

pip install auxknow

Usage Instructions

To begin using AuxKnow, ensure you have the required API keys (OPENAI_API_KEY and PERPLEXITY_API_KEY) configured in a .env file in your project’s root directory.

Here is an example of how to get started:

from auxknow import AuxKnow

# Initialize the Answer Engine
engine = AuxKnow()

# Ask a question
response = engine.ask("What is the capital of France?")

# Print the response and its citations
print(response.answer)
print(response.citations)

Use Cases

AuxKnow is designed to cater to a wide range of scenarios, including:

  • Vertical AI Agents: Build domain-specific AI agents tailored to industries such as healthcare, education, finance, or technology.

  • Advanced Q&A Systems: Integrate robust answering capabilities into your applications, powered by state-of-the-art Sonar models.

  • Custom User Experiences: Create personalized and immersive experiences with AuxKnow’s flexible configuration and session management features.

  • AI Infrastructure: Enhance your AI or LLM platform with AuxKnow to deliver best-in-class answering capabilities to your users.


Changelog

🚀 v0.0.16 - Design Improvements & Stability Fixes

  • 🎯 Improved overall design and stability of AuxKnow.
  • 🎯 Added improved intelligence to handle various edge cases, errors and operations.

🚀 v0.0.15 - Performance Logging

  • 🎯 Added enable_performance_logging configuration with configurable time units for performance logging.
  • 🎯 Enhanced design of verbose logging across the codebase.

🚀 v0.0.14 - Improved Memory Management

  • 🎯 Improved memory management in sessions.
  • 🎯 Codebase improvements, deprecated api_key in favour of perplexity_api_key.
  • 🛠 Centralized defaults and constants management.

🚀 v0.0.13 - Config Fixes

  • 🛠 Fix incorrect setting of config which was causing an error with set_config.

🚀 v0.0.12 - Fast Mode

  • 🎯 Fastest responses: The ask methods now support fast mode that can be enabled to receive the fastest possible responses at the cost of response quality and citation relevence.

🚀 v0.0.11 - Deep Research Mode

  • 🎯 Conduct Deep Research: The ask methods now support deep research mode that can be enabled with a flag.

🚀 v0.0.10 - Performance & Accuracy Boost

  • 🛠 Unbiased Reasoning Mode: Further refinements for enhanced neutrality.
  • Streaming Enhancements: Faster response times with optimized performance.
  • 🎯 Accuracy Improvements: Fine-tuned models for better contextual understanding.

🔍 v0.0.9 - DeepSeek R1 & Unbiased Reasoning

  • 🤖 Integrated DeepSeek R1: Uncensored responses with citations.
  • 🧠 Unbiased Reasoning Mode: Ensuring neutrality and factual accuracy.

✨ v0.0.8 - Auto Prompt Augmentation

  • 🚀 Auto Prompt Augmentation: Boosts response quality through automated prompt augmentation.

Contributing

We welcome contributions from developers around the globe. To get started, please review the Contributing Guidelines and submit your ideas, bug reports, or pull requests. Together, we can make AuxKnow even better.


License

AuxKnow is distributed under the AGPLv3 License. Refer to the LICENSE file for full details. Please also read the Terms of Use and Trademark.


Quote

"Knowledge is the foundation of progress." — The Hackers Playbook

Elevate your organization with AuxKnow and join us in redefining the future of AI-driven solutions.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

auxknow-0.0.16.tar.gz (41.2 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

auxknow-0.0.16-py3-none-any.whl (41.7 kB view details)

Uploaded Python 3

File details

Details for the file auxknow-0.0.16.tar.gz.

File metadata

  • Download URL: auxknow-0.0.16.tar.gz
  • Upload date:
  • Size: 41.2 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.1.0 CPython/3.12.4

File hashes

Hashes for auxknow-0.0.16.tar.gz
Algorithm Hash digest
SHA256 80b5b8240f49f10a531cf4721da9f4067bd6c43577fe3e4151a1931bd3dda5ff
MD5 9a8c196071fb51327b860b3a4dfc336e
BLAKE2b-256 c0c578cc0930869a1a99d3ef247233ad8ef0f61e72bf7804e3f01142f0f669d6

See more details on using hashes here.

File details

Details for the file auxknow-0.0.16-py3-none-any.whl.

File metadata

  • Download URL: auxknow-0.0.16-py3-none-any.whl
  • Upload date:
  • Size: 41.7 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.1.0 CPython/3.12.4

File hashes

Hashes for auxknow-0.0.16-py3-none-any.whl
Algorithm Hash digest
SHA256 2ed4c00e6cd53a19603115216eab32b9f07a1705a3d5c22110d6d7bbc4e1555e
MD5 22b3a3a041da86245ec3cf981604c4eb
BLAKE2b-256 0b887837108347d3831d4720294487ee12cc4195fc0e2771228d14094c16ada1

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page