Package com.tnsai.llm.providers

package com.tnsai.llm.providers

LLM provider implementations for various AI services.

This package contains concrete implementations of the LLMClient interface for different LLM providers. Each client handles provider-specific API details while exposing a unified interface.

Available Providers

Client	Provider	API Key Env Var
OpenAIClient	OpenAI	OPENAI_API_KEY
AzureOpenAIClient	Azure OpenAI	AZURE_OPENAI_API_KEY
AnthropicClient	Anthropic	ANTHROPIC_API_KEY
GeminiClient	Google Gemini	GEMINI_API_KEY
OllamaClient	Ollama (local)	N/A
GroqClient	Groq	GROQ_API_KEY
OpenRouterClient	OpenRouter	OPENROUTER_API_KEY
MistralClient	Mistral AI	MISTRAL_API_KEY
BedrockClient	AWS Bedrock	AWS credentials

Common Features

• Retry Logic: Exponential backoff on rate limits (429) and server errors (5xx)
• Tool Calling: Function/tool support where provider allows
• Configurable: Model, temperature, max tokens, etc.
• Error Handling: Provider-specific error mapping to common exceptions

Usage Examples

OpenAI

 LLMClient client = new OpenAIClient("gpt-4");
 String response = client.chat("Explain quantum computing");
 

Anthropic

 LLMClient client = new AnthropicClient("claude-3-opus-20240229");
 String response = client.chat("Write a haiku about programming");
 

Ollama (Local)

 LLMClient client = new OllamaClient("llama2");
 // Uses localhost:11434 by default
 String response = client.chat("What is 2+2?");
 

Gemini with Images

 GeminiClient client = new GeminiClient("gemini-pro-vision");
 List<ContentPart> parts = List.of(
     new TextPart("Describe this image"),
     ImagePart.fromBase64(imageData, "image/jpeg")
 );
 String response = client.chat(parts);
 

Provider Selection Guide

• Best quality: GPT-4, Claude 3 Opus, Gemini Ultra
• Best speed: Groq (Llama/Mixtral), GPT-3.5-turbo
• Best value: OpenRouter, Ollama (free local)
• Enterprise: Azure OpenAI, AWS Bedrock
• Privacy: Ollama (local), Azure (private deployment)

See Also:

• LLMClient

Related Packages

Package	Description
com.tnsai.llm

Class	Description
AbstractLLMClient	Abstract base class for LLM client implementations.
AbstractOpenAICompatibleClient	Base class for providers that speak the OpenAI /chat/completions wire format verbatim — same request body, same choices[].message response, same choices[].delta streaming, Bearer auth.
AnthropicClient	Anthropic Claude LLM client implementation.
AnthropicClient.Builder	Builder for creating AnthropicClient instances.
AnthropicToolConverter	Translates the OpenAI-flavoured intermediate tool schema into Anthropic's Messages-API tool shape.
AzureOpenAIClient	Azure OpenAI Service client implementation.
AzureOpenAIClient.AuthType	Authentication type for Azure OpenAI.
AzureOpenAIClient.AzureContentFilterException	Exception thrown when Azure OpenAI content filter rejects the request.
BedrockClient	AWS Bedrock LLM client implementation.
CerebrasClient	Cerebras LLM client.
CohereClient	Cohere LLM client implementation.
CohereToolConverter	Translates the OpenAI-flavoured intermediate tool schema into Cohere's chat-API tool shape.
DatabricksClient	Databricks Mosaic AI Model Serving LLM client.
DeepInfraClient	DeepInfra LLM client.
DeepSeekClient	DeepSeek LLM client.
FireworksAIClient	Fireworks.ai LLM client.
GeminiClient	Google Gemini LLM client with multimodal support (text, image, audio, video).
GroqClient	Groq LLM client implementation.
HuggingFaceClient	HuggingFace Inference API client using OpenAI-compatible endpoint.
LlamaCppServerClient	llama.cpp server LLM client.
LMStudioClient	LM Studio LLM client.
MiniMaxClient	MiniMax LLM client implementation.
MiniMaxClient.Builder
MistralClient	Mistral AI LLM client implementation.
NvidiaNIMClient	NVIDIA NIM (NVIDIA Inference Microservices) LLM client.
OllamaClient	Ollama LLM client implementation for local and cloud deployments.
OpenAIClient	OpenAI LLM client implementation.
OpenRouterClient	OpenRouter LLM client implementation.
PerplexityClient	Perplexity LLM client.
QwenCloudClient	Alibaba Qwen Cloud (DashScope) LLM client.
ReplicateClient	Replicate LLM client.
TencentHunyuanClient	Tencent Hunyuan LLM client.
TogetherAIClient	Together.ai LLM client.
VertexAIClient	Google Cloud Vertex AI LLM client (enterprise Gemini).
VLLMClient	vLLM LLM client.
WatsonxClient	IBM watsonx.ai LLM client.
XAIGrokClient	xAI Grok LLM client.
YiClient	01.AI Yi LLM client.
ZhipuAIClient	ZhipuAI (Z.AI/Zhipu) LLM client implementation.
ZhipuAIClient.Builder