Request a Demo
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.

Release Notes - AltrumAI Platform 1.1

Features and Improvement

Production release of Python SDK

Python SDK designed to streamline integration, improve the developer experience, and encourage the adoption of AI Guardrails among Python developers and their Gen AI workflows.

Official release of the AltrumAI Website

First working version of AltrumAI website. The new website provides comprehensive information about platform's features, benefits, and capabilities, along with detailed documentation, tutorials, and resources to help users get started.

Implemented new Context guardrails

AltrumAI platform now includes contextual guardrails. These guardrails ensure that generated content is appropriate and stays relevant by maintaining focus on specified topics and avoiding sensitive subjects.

  1. Restrict to Topic
  2. Sensitive Topic
  3. Competitor Mention
  4. Drugs Mention
  5. Unusual Prompt
  6. QA Relevance
  7. Response Evaluation
  8. Politeness Check
  9. LLM Critic
  10. Gibberish Text
  11. NSFW Text
  12. Financial Tone
  13. Profanity
  14. Secrets Present

Enhanced Data Privacy guardrails

Enhanced Data Privacy guardrails by adding new regulatory groups and relevant entities in addition to GDPR, HIPAA and Custom entities

  1. PCI DSS
  2. Quebec Privacy Act
  3. CPRA
  4. APPI
  5. PII
  6. PHI

Implemented Text to Audio & Audio to Text models

Enhanced modalities of text-to-audio and audio-to-text models to AltrumAI Platform. These new features enable seamless conversion between text and audio, enhancing accessibility and expanding the range of applications for our safe AI solutions.

Bug Fixes

  1. Data Privacy API now accepts custom entities as per users Policies.
  1. The Ethical AI layer API now supports 30000 characters as input from the API users.

Known Issues

  1. (Non-functional) HTTP Error 504 error for cold starts in the sandbox workspace.
  1. (Non-functional) HTTP 500 internal server error when using model as "llama-3-8b-instruct" for chat completions.
  1. (Non-functional) HTTP 500 internal server error while using "mistral-7b-chat" when giving input in the range of 5k to 30k characters in multiple prompts.
  1. (Public Documentation) API Documentation to be updated for Moderations guardrail in case of 422 error.
  1. (Public Documentation) API Documentation to be updated with appropriate request and response details for GET ALL MODELS endpoint.