pipecat-costing-sdk 0.0.5

Get a costing of voice ai pipeline. when call ends

Pipecat Costing SDK

A Python SDK for calculating costs associated with voice AI pipelines using Pipecat metrics. This library helps track and estimate costs for LLM, TTS, and STT usage in real-time voice applications.

Features

• Tracks LLM token usage (prompt, completion, cache read, cache creation)
• Monitors TTS character count
• Estimates session duration for STT costs
• Provides detailed cost breakdowns and grand totals
• Integrates seamlessly with Pipecat's frame processing pipeline

Installation

Install the package from PyPI:

pip install pipecat-costing-sdk

Usage

from pipecat_costing_sdk import CostLogger

# Initialize the cost logger with pricing rates
cost_logger = CostLogger(
    PRICE_PER_1K_PROMPT_TOKENS=0.0015,
    PRICE_PER_1K_COMPLETION_TOKENS=0.002,
    PRICE_PER_1K_CACHE_READ_TOKENS=0.00075,
    PRICE_PER_1K_CACHE_CREATION_TOKENS=0.00125,
    PRICE_PER_1K_TTS_CHARS=0.015,
    PRICE_PER_STT_MINUTE=0.011
)

# Integrate into your Pipecat pipeline
# (Assuming you have a pipeline set up with metrics frames)

# After processing, get the cost report
usage, costs = cost_logger.get_cost_report
print("Usage:", usage)
print("Costs:", costs)

API Reference

CostLogger

• __init__(PRICE_PER_1K_PROMPT_TOKENS: float, PRICE_PER_1K_COMPLETION_TOKENS: float, PRICE_PER_1K_CACHE_READ_TOKENS: float, PRICE_PER_1K_CACHE_CREATION_TOKENS: float, PRICE_PER_1K_TTS_CHARS: float, PRICE_PER_STT_MINUTE: float): Initialize with pricing rates per unit
• get_cost_report: Property that returns a tuple of (usage_dict, cost_dict)

Usage Dictionary

• prompt_tokens: Total prompt tokens used
• completion_tokens: Total completion tokens used
• cache_read_tokens: Tokens read from cache
• cache_creation_tokens: Tokens used to create cache
• tts_characters: Total characters synthesized
• session_seconds: Total session duration in seconds

Cost Dictionary

• llm_cost_prompt: Cost for prompt tokens
• llm_cost_completion: Cost for completion tokens
• llm_cost_cache_read: Cost for cache read tokens
• llm_total_cost: Total LLM cost
• tts_cost: Cost for TTS
• stt_cost: Cost for STT
• grand_total: Total cost for the session

Dependencies

• pipecat

License

This project is licensed under the MIT License - see the LICENSE file for details.

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

Support

If you encounter any issues or have questions, please open an issue on GitHub.