Details

Summary

Added support for tracking OpenAI Text-to-Speech (TTS) models (tts-1, tts-1-hd) usage and costs within the Opik platform.

Key Changes

• TTS Tracking: enabled tracking for audio.speech.create and audio.speech.with_streaming_response.create.
• Cost Calculation: Implemented character-based cost tracking for TTS models as per OpenAI pricing.
• Audio Logging: Generated audio is now automatically captured and logged as an attachment to the trace.

Change checklist

• User facing
• Documentation update

Issues

#2202 /claim #2202

• Resolves #2202
• OPIK-

Testing

Tested locally

Documentation