ai-infra 0.1.109

Infrastructure for efficient and scalable AI applications.

ai-infra

Production-ready Python SDK for building AI applications with LLMs, agents, and multimodal capabilities.

ai-infra provides clean interfaces for chat, agents, embeddings, voice, and image generation across 10+ providers—all with zero-config defaults.

✨ Features

• LLM: Chat, structured output, streaming, retries, multi-turn conversations
• Agents: Tool calling, human-in-the-loop, provider fallbacks, autonomous deep mode
• Graph: LangGraph workflows with typed state and conditional branching
• Embeddings & RAG: Vector storage, document retrieval, multiple backends
• Multimodal: Text-to-speech, speech-to-text, vision, realtime voice
• Image Generation: DALL-E, Imagen, Stability AI, Replicate
• MCP: Model Context Protocol client/server, OpenAPI→MCP conversion

🚀 Quick Start

5 lines to your first chat:

from ai_infra import LLM

llm = LLM()  # Auto-detects configured provider
response = llm.chat("What is the capital of France?")
print(response)

With tools (agent):

from ai_infra import Agent

def get_weather(city: str) -> str:
    """Get weather for a city."""
    return f"Weather in {city}: 72°F, sunny"

agent = Agent(tools=[get_weather])
result = agent.run("What's the weather in Tokyo?")
print(result)

📦 Installation

Python: 3.11 – 3.13

# Using pip
pip install ai-infra

# Using Poetry (development)
poetry install
poetry shell

🔑 Provider Setup

Set API keys for the providers you want to use:

# Required: At least one chat provider
export OPENAI_API_KEY=sk-...
export ANTHROPIC_API_KEY=sk-ant-...
export GOOGLE_API_KEY=...
export XAI_API_KEY=...

# Optional: Specialized providers
export ELEVENLABS_API_KEY=...     # TTS
export DEEPGRAM_API_KEY=...       # STT
export STABILITY_API_KEY=...      # Image generation
export REPLICATE_API_TOKEN=...    # Image generation
export VOYAGE_API_KEY=...         # Embeddings
export COHERE_API_KEY=...         # Embeddings

🔌 Supported Providers

Provider	Chat	Embeddings	TTS	STT	ImageGen	Realtime
OpenAI	✅	✅	✅	✅	✅	✅
Anthropic	✅	-	-	-	-	-
Google	✅	✅	✅	✅	✅	✅
xAI	✅	-	-	-	-	-
ElevenLabs	-	-	✅	-	-	-
Deepgram	-	-	-	✅	-	-
Stability	-	-	-	-	✅	-
Replicate	-	-	-	-	✅	-
Voyage	-	✅	-	-	-	-
Cohere	-	✅	-	-	-	-

📚 Documentation

Full documentation is in the docs/ folder:

Section	Description
Getting Started	Installation, API keys, first example
Core Modules	LLM, Agent, Graph, Providers
Multimodal	TTS, STT, Vision, Realtime Voice
Embeddings & RAG	Embeddings, VectorStore, Retriever
Tools	Schema tools, progress streaming
MCP	Model Context Protocol client/server
Advanced Features	Personas, Replay, Workspace, Deep Agent
Image Generation	DALL-E, Imagen, Stability, Replicate
Infrastructure	Errors, Logging, Tracing, Callbacks
CLI Reference	Command-line interface

📁 Module Overview

Module	Description
ai_infra.llm	LLM chat, agents, structured output, streaming
ai_infra.graph	LangGraph workflows with typed state
ai_infra.mcp	MCP client/server, OpenAPI→MCP conversion
ai_infra.embeddings	Text embeddings across providers
ai_infra.retriever	RAG with multiple vector store backends
ai_infra.imagegen	Image generation (DALL-E, Stability, etc.)
ai_infra.providers	Centralized provider registry

🧪 Examples

See the examples/ folder for runnable scripts:

# LLM chat
python -c "from ai_infra.llm.examples.02_llm_chat_basic import main; main()"

# Agent with tools
python -c "from ai_infra.llm.examples.01_agent_basic import main; main()"

# Graph workflow
python -c "from ai_infra.graph.examples.01_graph_basic import main; main()"

# MCP client
python -m ai_infra.mcp.examples.01_mcps

🛠️ Development

# Install dev dependencies
poetry install

# Run tests
pytest -q

# Lint
ruff check src tests

# Type check
mypy src

# Format
ruff format

📄 License

MIT