AI Providers

Integrate AI providers like OpenAI, Claude, and Gemini to build smarter, modular workflows effortlessly.

Polylith empowers developers to integrate seamlessly with top AI providers, enabling everything from text generation and chat agents to automation workflows and real-time analysis. With Polylith’s modular architecture, you can connect these providers effortlessly, swap models in seconds, and unlock the full potential of modern AI.

Supported Providers

Polylith currently supports:

OpenAI Use models like GPT-3.5 and GPT-4 for text generation, summarization, question-answering, and creative tasks.
Claude by Anthropic Build conversational agents or reasoning workflows with Claude v1 and Claude v2 for natural dialogue and decision-making.
Gemini by Google Integrate Google’s multimodal AI models to process text, vision, and structured data for more complex workflows.
Grok by xAI Leverage Grok for AI reasoning and automation, especially in logic-driven tools.
Perplexity AI Access real-time search, summarization, and data analysis tools with Perplexity AI to integrate up-to-date information.

How Polylith Makes It Easy

Polylith’s Input, Processing, and Output Blocks simplify connecting AI providers into your workflows. Each provider can be added in just a few lines of code, and switching between them is frictionless.

Example 1: Text Generation with OpenAI GPT-4

This example demonstrates a text summarization pipeline using OpenAI:

Copy

pythonCopy codefrom polylith import InputBlock, ProcessingBlock, OutputBlock  

# Input Block: Capture the text you want summarized  
input_block = InputBlock("user_input", prompt="Summarize the following: Polylith simplifies AI development workflows.")  

# Processing Block: Use OpenAI's GPT-4 API for text generation  
processing_block = ProcessingBlock("openai", model="gpt-4", api_key="YOUR_OPENAI_API_KEY")  

# Output Block: Send the AI response to the console  
output_block = OutputBlock("console")  

# Workflow: Connect the blocks and run the pipeline  
workflow = input_block >> processing_block >> output_block  
workflow.run()

Example 2: Switching to Claude with Minimal Changes

To use Anthropic’s Claude instead of OpenAI, simply change the provider in the ProcessingBlock:

Copy

pythonCopy code# Processing Block: Switch from OpenAI to Claude  
processing_block = ProcessingBlock("claude", model="claude-v2", api_key="YOUR_CLAUDE_API_KEY")

Everything else in your workflow remains intact. Polylith’s modular approach allows you to experiment and optimize without rewriting entire codebases.

Example 3: Building a Chat Agent

Here’s how you can create a chat agent that listens to input, generates responses using Gemini, and outputs back to a user interface:

Copy

pythonCopy codefrom polylith import InputBlock, ProcessingBlock, OutputBlock  

# Input Block: Capture user messages (e.g., from Discord or Console)  
input_block = InputBlock("discord", token="YOUR_DISCORD_TOKEN", channel_id="1234567890")  

# Processing Block: Use Google's Gemini API to process the message  
processing_block = ProcessingBlock("gemini", model="gemini-1", api_key="YOUR_GEMINI_API_KEY")  

# Output Block: Send the AI-generated response back to the channel  
output_block = OutputBlock("discord_response", token="YOUR_DISCORD_TOKEN")  

# Workflow: Chain the blocks together  
workflow = input_block >> processing_block >> output_block  
workflow.run()

Example 4: Integrating Perplexity AI for Real-Time Search

Combine real-time search capabilities from Perplexity AI with Polylith workflows:

Copy

pythonCopy codefrom polylith import InputBlock, ProcessingBlock, OutputBlock  

# Input Block: A search query provided by the user  
input_block = InputBlock("user_input", prompt="What is Polylith?")  

# Processing Block: Perplexity AI for real-time search and summarization  
processing_block = ProcessingBlock("perplexity", api_key="YOUR_PERPLEXITY_API_KEY")  

# Output Block: Print the results to the console  
output_block = OutputBlock("console")  

# Workflow: Link the blocks and execute  
workflow = input_block >> processing_block >> output_block  
workflow.run()

Why Modular AI Providers Matter

Experimentation Made Easy Test multiple providers (OpenAI, Claude, Gemini) within the same workflow to find the best solution for your task.
Cost and Performance Optimization Switch between providers depending on cost efficiency, performance needs, or API capabilities.
Simplified Integration With Polylith, complex provider APIs are abstracted into clean, reusable blocks, making it easy to focus on your logic rather than boilerplate code.

Real-World Use Cases

Content Generation: Generate marketing copy, blog posts, or creative content using OpenAI or Claude.
Conversational AI: Deploy chatbots and virtual assistants for Discord, Telegram, or web apps.
Automated Summarization: Summarize long documents, research papers, or articles.
Knowledge Agents: Combine Perplexity AI for real-time search with reasoning workflows powered by Grok.

Conclusion

Polylith’s AI Provider integrations give you the freedom to experiment, scale, and optimize your AI tools without friction. Whether you’re building real-time chat agents, automating content, or enhancing decision-making workflows, Polylith makes it easy to leverage the best AI models available today.

PreviousWhat is Polylith?NextVector Stores

Last updated 1 month ago

pythonCopy codefrom polylith import InputBlock, ProcessingBlock, OutputBlock # Input Block: Capture the text you want summarized input_block = InputBlock("user_input", prompt="Summarize the following: Polylith simplifies AI development workflows.") # Processing Block: Use OpenAI's GPT-4 API for text generation processing_block = ProcessingBlock("openai", model="gpt-4", api_key="YOUR_OPENAI_API_KEY") # Output Block: Send the AI response to the console output_block = OutputBlock("console") # Workflow: Connect the blocks and run the pipeline workflow = input_block >> processing_block >> output_block workflow.run()

pythonCopy codefrom polylith import InputBlock, ProcessingBlock, OutputBlock # Input Block: Capture user messages (e.g., from Discord or Console) input_block = InputBlock("discord", token="YOUR_DISCORD_TOKEN", channel_id="1234567890") # Processing Block: Use Google's Gemini API to process the message processing_block = ProcessingBlock("gemini", model="gemini-1", api_key="YOUR_GEMINI_API_KEY") # Output Block: Send the AI-generated response back to the channel output_block = OutputBlock("discord_response", token="YOUR_DISCORD_TOKEN") # Workflow: Chain the blocks together workflow = input_block >> processing_block >> output_block workflow.run()

pythonCopy codefrom polylith import InputBlock, ProcessingBlock, OutputBlock # Input Block: A search query provided by the user input_block = InputBlock("user_input", prompt="What is Polylith?") # Processing Block: Perplexity AI for real-time search and summarization processing_block = ProcessingBlock("perplexity", api_key="YOUR_PERPLEXITY_API_KEY") # Output Block: Print the results to the console output_block = OutputBlock("console") # Workflow: Link the blocks and execute workflow = input_block >> processing_block >> output_block workflow.run()