Streamline your AI workflow with a unified, elegant interface for multiple AI providers

Hex.pm Documentation License

Working with different AI models shouldn't feel like herding cats. AiFlow provides a consistent, developer-friendly interface that makes integrating AI into your Elixir applications a breeze. Start with Ollama today, with more providers coming soon.

# Ask any question - it's that simple!
{:ok, response} = AiFlow.Ollama.query("Explain quantum computing in simple terms", "llama3.1")

One interface, multiple AI providers. Switch between services without rewriting your code.

Built-in error handling, debugging tools, and comprehensive testing.

• 🧠 Model Management: List, create, copy, delete, pull, and push models
• 💬 Smart Chat Sessions: Persistent chat history with automatic context management
• ✍️ Text Generation: Powerful prompt completion with customizable parameters
• 🔍 Embeddings: Generate vector embeddings for semantic search and ML tasks
• 🔄 Blob Operations: Efficient model file management
• 🛡️ Robust Error Handling: Comprehensive error management with bang (!) versions
• 🐛 Advanced Debugging: Built-in tools for troubleshooting and development

Add ai_flow to your list of dependencies in mix.exs:

def deps do
 [
 {:ai_flow, "~> 0.1.0"}
 ]
end

# Quick start with defaults
{:ok, pid} = AiFlow.Ollama.start_link()
# Or customize your setup
{:ok, pid} = AiFlow.Ollama.start_link(
 hostname: "localhost",
 port: 11434,
 timeout: 60_000
)

# Simple question
{:ok, response} = AiFlow.Ollama.query("Why is the sky blue?", "llama3.1")
# Interactive chat
{:ok, response} = AiFlow.Ollama.chat("Hello!", "chat_session_1", "user_123", "llama3.1")
{:ok, response} = AiFlow.Ollama.chat("Tell me more about that", "chat_session_1", "user_123", "llama3.1")

# Generate embeddings for semantic search
{:ok, embeddings} = AiFlow.Ollama.generate_embeddings([
 "The cat sat on the mat",
 "A feline rested on the rug"
])
# Manage your models
{:ok, models} = AiFlow.Ollama.list_models()
{:ok, :success} = AiFlow.Ollama.create_model("my-custom-model", "llama3.1", "You are a helpful coding assistant.")

Work with AI models intuitively:

• AiFlow.Ollama.list_models() - Discover available models
• AiFlow.Ollama.query() - Ask questions to any model
• AiFlow.Ollama.chat() - Engage in persistent conversations

# Everything you need to manage AI models
AiFlow.Ollama.list_models()
AiFlow.Ollama.create_model("my-model", "base-model", "system prompt")
AiFlow.Ollama.copy_model("original", "backup")
AiFlow.Ollama.delete_model("old-model")
AiFlow.Ollama.pull_model("new-model")
AiFlow.Ollama.push_model("my-model:latest")

Flexible configuration for any environment:

# Application-wide configuration
config :ai_flow, AiFlow.Ollama,
 hostname: "localhost",
 port: 11434,
 timeout: 60_000
# Or per-instance configuration
{:ok, pid} = AiFlow.Ollama.start_link(
 hostname: "production-ai.internal",
 port: 11434,
 timeout: 120_000
)

AiFlow is just getting started! Upcoming integrations include:

• 🐝 Bumblebee Integration: Hugging Face models support
• ☁️ Cloud AI Providers: OpenAI, Anthropic, Google AI
• 📦 Model Registry: Centralized model management
• ⚡ Performance Optimizations: Caching and batching

We love contributions! Here's how to get started:

1. Fork the repository
2. Create your feature branch (git checkout -b feature/AmazingFeature)
3. Commit your changes (git commit -m 'Add some AmazingFeature')
4. Push to the branch (git push origin feature/AmazingFeature)
5. Open a Pull Request

Full API documentation is available at HexDocs.

Distributed under the MIT License. See LICENSE for more information.

• Found a bug? Open an issue
• Have a feature request? We'd love to hear it!
• Questions? Check out the documentation or open a discussion

Made with ❤️ for the Elixir community