Skip to content

Cognix v0.1.3 - OpenRouter Integration

Choose a tag to compare

View all tags
@cognix-dev cognix-dev released this 01 Sep 01:27
· 78 commits to main since this release
v0.1.3
11b061d

Cognix v0.1.3 Release Notes

Released: September 1, 2025

馃殌 New Features

OpenRouter Integration

  • Multiple AI models with single API key - Access Google Gemini, Microsoft Phi, Meta Llama, and many more models through OpenRouter
  • Free model support - Use powerful models like google/gemini-2.0-flash-exp:free at no cost
  • Seamless provider switching - Switch between OpenRouter, OpenAI, and Anthropic without configuration changes

Enhanced Error Handling

  • Smart error recovery - Automatic fallback suggestions for common API issues
  • User-friendly error messages - Clear guidance for 402 (insufficient credits), 429 (rate limited), and 401 (invalid key) errors
  • Provider-specific troubleshooting - Contextual help based on your current API configuration

馃敡 Improvements

Model Management

  • Improved model detection - Better automatic detection of available models based on configured API keys
  • Enhanced model switching - More reliable switching between different providers and models
  • Provider identification - Clear display of which provider is currently active (Direct vs OpenRouter)

User Experience

  • Better setup guidance - Clearer documentation for API key configuration
  • Improved error context - More helpful error messages with specific next steps
  • Configuration validation - Automatic detection and resolution of common setup issues

馃悰 Bug Fixes

  • Fixed model availability detection when switching providers
  • Resolved OpenRouter API key handling in multi-provider setups
  • Corrected provider display information for OpenRouter models
  • Fixed model switching behavior when OPENAI_BASE_URL is set

馃摎 Documentation Updates

  • Added comprehensive OpenRouter setup guide
  • Updated .env.example with recommended configuration patterns
  • Enhanced troubleshooting section with provider-specific solutions
  • Added free model discovery guide

馃攧 Migration Guide

For Existing Users

Your existing configuration will continue to work without changes. To take advantage of OpenRouter features:

  1. Optional OpenRouter setup:

    OPENAI_API_KEY=sk-or-v1-your_openrouter_key
OPENAI_BASE_URL=https://openrouter.ai/api/v1

  2. Access free models:

    cognix> /model google/gemini-2.0-flash-exp:free

Configuration Patterns

  • Anthropic only: Set ANTHROPIC_API_KEY
  • OpenAI only: Set OPENAI_API_KEY
  • OpenRouter: Set OPENAI_API_KEY (with OpenRouter key) + OPENAI_BASE_URL
  • Mixed setup: Any combination of the above

馃幆 What's Coming Next

v0.2.0 Preview

  • Memory management improvements
  • Enhanced code analysis features
  • Performance optimizations
  • Additional model providers

馃搵 Technical Details

API Compatibility

  • Maintains backward compatibility with existing configurations
  • Supports all previous model names and aliases
  • No breaking changes to CLI commands or session format

Provider Support

  • Anthropic: Claude 4 series, Claude 3.7, Claude 3.5
  • OpenAI: GPT-4o series, GPT-4, GPT-3.5-turbo
  • OpenRouter: 100+ models from multiple providers including free options

System Requirements

  • Python 3.8+
  • Internet connection for API access
  • At least one configured API provider

馃檹 Acknowledgments

Thanks to the community for feedback on multi-provider support and error handling improvements. Special recognition for users who reported model switching issues that led to this enhanced implementation.

Install or upgrade: pip install --upgrade cognix

Get started: https://github.com/cognix-dev/cognix#quick-start

Report issues: https://github.com/cognix-dev/cognix/issues

Assets 2
Loading