Feature Request: Add OpenRouter embeddings support

[leex279]

Feature Request: OpenRouter Embeddings Support

Description

OpenRouter now supports embeddings models, which would allow Archon users to leverage a variety of embedding providers through a single API endpoint.

Background

OpenRouter has added support for embeddings models: https://openrouter.ai/models?fmt=cards&output_modalities=embeddings

This includes models from various providers accessible through OpenRouter's unified API.

Benefits

• Unified API: Users can access multiple embedding providers through one API key
• Model Variety: Access to different embedding models (e.g., text-embedding-3-small, text-embedding-3-large, etc.)
• Simplified Setup: Single API configuration instead of managing multiple provider credentials
• Consistent Interface: OpenRouter provides a unified interface similar to OpenAI's API

Technical Considerations

• OpenRouter uses an OpenAI-compatible API for embeddings
• Should integrate with existing embedding provider infrastructure in python/src/server/services/embeddings/
• Configuration would require OPENROUTER_API_KEY environment variable
• Base URL would be https://openrouter.ai/api/v1

Implementation Scope

• Add OpenRouter as an embedding provider option
• Update environment configuration documentation
• Add OpenRouter-specific models to the model selection UI
• Ensure proper error handling and API compatibility

Related

This would complement the existing OpenAI, Cohere, and other embedding provider support already in Archon.