Anannas AI Integration for Pipecat

An OpenAI-compatible LLM service integration for Pipecat that provides access to 500+ models through Anannas AI's unified gateway.

About Anannas AI

Anannas AI is a unified inference gateway that provides access to 500+ models from OpenAI, Anthropic, Mistral, Gemini, DeepSeek, and other providers through a single OpenAI-compatible API. Key features include:

• Unified API: Access 500+ models through one consistent interface
• Built-in Observability: Cache hit rate analytics, token-level metrics, tool/function call analytics, and model efficiency scoring
• Smart Routing: Automatic provider health monitoring and failover with ~0.48ms overhead
• Cost Optimization: 5% markup with transparent pricing
• Enterprise Ready: BYOK (Bring Your Own Key) support for secure deployments
• Production Proven: Powering deployments with 100k+ requests and 1B+ tokens processed

Installation

Install the package from PyPI:

pip install pipecat-anannas

Or install from source:

git clone https://github.com/upsurgeio/anannas-pipecat-integration.git
cd anannas-pipecat-integration
pip install -e .

Quick Start

import os
from pipecat_anannas import AnannasLLMService
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.task import PipelineTask

# Initialize Anannas LLM service
llm = AnannasLLMService(
    api_key=os.getenv("ANANNAS_API_KEY"),
    model="gpt-4o"  # or any supported model
)

# Use in your Pipecat pipeline
pipeline = Pipeline([
    # ... your pipeline components
    llm,
    # ... more components
])

Supported Models

Anannas AI provides access to models from multiple providers:

• OpenAI: gpt-5-codex, gpt-5-pro, gpt-5-chat, and more
• Anthropic: claude-opus-4.1, claude-opus-4, claude-sonnet-4.5
• Mistral: magistral-medium-2506, codestral-2508, devstral-medium
• Google: gemini-2.5-flash-image-preview, gemma-2-27b-it
• DeepSeek: deepseek-chat, deepseek-r1
• And 500+ more models...

Usage with Pipecat Pipeline

Basic Example

import os
from dotenv import load_dotenv

from pipecat_anannas import AnannasLLMService
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.task import PipelineTask
from pipecat.services.deepgram.stt import DeepgramSTTService
from pipecat.services.azure.tts import AzureTTSService

load_dotenv()

# Create services
stt = DeepgramSTTService(api_key=os.getenv("DEEPGRAM_API_KEY"))
llm = AnannasLLMService(api_key=os.getenv("ANANNAS_API_KEY"), model="gpt-4o")
tts = AzureTTSService(
    api_key=os.getenv("AZURE_SPEECH_API_KEY"),
    region=os.getenv("AZURE_SPEECH_REGION")
)

# Build pipeline
pipeline = Pipeline([stt, llm, tts])
task = PipelineTask(pipeline)

Function Calling Example

See the complete function calling example in examples/function_calling_example.py. This demonstrates:

• Setting up Anannas LLM service
• Registering function handlers
• Using function calling in a voice conversation
• Streaming responses

To run the example:

# Set required environment variables
export ANANNAS_API_KEY=your_api_key
export DEEPGRAM_API_KEY=your_deepgram_key
export AZURE_SPEECH_API_KEY=your_azure_key
export AZURE_SPEECH_REGION=your_azure_region

# Run the example
python examples/function_calling_example.py

Configuration

Environment Variables

• ANANNAS_API_KEY: Your Anannas AI API key (required)

Service Parameters

AnannasLLMService(
    api_key: Optional[str] = None,  # API key or use ANANNAS_API_KEY env var
    model: str = "gpt-4o",           # Model to use
    base_url: str = "https://api.anannas.ai/v1",  # API endpoint
    **kwargs                          # Additional OpenAI-compatible parameters
)

Observability

Anannas AI provides built-in observability through its dashboard:

• Cache Analytics: Monitor cache hit rates to optimize costs
• Token Metrics: Track token usage across models and requests
• Function Call Analytics: Analyze tool/function call patterns
• Model Efficiency: Compare performance across different models
• Provider Health: Real-time monitoring of provider availability

Access your dashboard at: https://anannas.ai/dashboard

Enterprise Features

BYOK (Bring Your Own Key)

For enterprise deployments requiring direct API keys to providers:

• Use your own OpenAI, Anthropic, or other provider API keys
• Maintain full control over credentials
• Transparent cost tracking

Learn more: https://docs.anannas.ai/UseCases/BYOK

Pipecat Compatibility

Tested with Pipecat: v0.0.86+

Links

• Anannas AI Homepage: https://anannas.ai
• Documentation: https://docs.anannas.ai
• Quickstart Guide: https://docs.anannas.ai/quickstart
• Enterprise: https://anannas.ai/enterprise
• BYOK Documentation: https://docs.anannas.ai/UseCases/BYOK
• Pipecat Documentation: https://docs.pipecat.ai

Support

• Issues: https://github.com/upsurgeio/anannas-pipecat-integration/issues
• Pipecat Discord: https://discord.gg/pipecat (#community-integrations channel)
• Anannas Support: [email protected]

Maintainer

This integration is maintained by the Anannas AI team.

License

BSD 2-Clause License - see LICENSE file for details.

This license is compatible with Pipecat's BSD 2-Clause License, ensuring seamless integration.

Contributing

Contributions are welcome! Please feel free to submit issues or pull requests.

Changelog

See CHANGELOG.md for version history and updates.