Supported Providers

Supported AI Providers

Rawi CLI supports 11 major AI providers, giving you access to dozens of models from leading companies. Whether you need cloud-based power or local privacy, there’s a provider configuration for your use case.

Provider Overview

Cloud Providers
Local Providers

Commercial cloud providers with API keys:

OpenAI - GPT-4o, GPT-4, O1 models
Anthropic - Claude 3.5 Sonnet, Claude 4
Google - Gemini 2.0 Flash, Gemini 1.5 Pro
DeepSeek - Cost-effective reasoning models
Mistral - European AI with multilingual support
Cerebras - Ultra-fast inference with WSE technology
Azure OpenAI - Enterprise OpenAI models
Amazon Bedrock - Multi-vendor model access
xAI - Grok models from Elon Musk’s company

Quick Provider Comparison

Provider	Best For	Key Models	Setup Difficulty
OpenAI	General AI tasks, coding	GPT-4o, O1	Easy
Anthropic	Long contexts, analysis	Claude 3.5 Sonnet	Easy
Google	Multimodal tasks	Gemini 2.0 Flash	Easy
DeepSeek	Cost-effective reasoning	DeepSeek Chat	Easy
Mistral	European AI, multilingual	Mistral Large	Easy
Cerebras	Ultra-fast inference	Llama 3.1 70B	Easy
Ollama	Privacy, local use	Llama 3.2, Mistral	Medium
LM Studio	Local model management	Ollama models	Easy
Azure	Enterprise environments	GPT-4 (Enterprise)	Medium
Bedrock	AWS ecosystem	Claude, Llama, Titan	Hard
xAI	Experimental features	Grok-2	Easy

Provider Details

OpenAI

The most popular provider with cutting-edge models including the latest GPT-4o and reasoning models.

Available Models:

gpt-4o - Latest multimodal model (recommended)
gpt-4o-mini - Faster, cost-effective version
gpt-4 - Previous generation flagship
gpt-3.5-turbo - Fast and affordable
o1-preview - Advanced reasoning model
o1-mini - Compact reasoning model

Configuration:

# Basic setup
rawi configure --provider openai --model gpt-4o --api-key sk-your-key

# With custom settings
rawi configure --provider openai \
  --model gpt-4o \
  --api-key sk-your-key \
  --temperature 0.7 \
  --max-tokens 2048

Use Cases:

General purpose AI tasks
Code generation and review
Writing and content creation
Data analysis and reasoning

Anthropic (Claude)

Anthropic’s Claude models excel at long-form analysis, safety, and nuanced reasoning.

Available Models:

claude-3-5-sonnet-20241022 - Latest and most capable (recommended)
claude-3-5-haiku-20241022 - Fast and cost-effective
claude-3-opus-20240229 - Previous flagship model

Configuration:

# Basic setup
rawi configure --provider anthropic \
  --model claude-3-5-sonnet-20241022 \
  --api-key sk-ant-your-key

# For long-form analysis
rawi configure --provider anthropic \
  --model claude-3-5-sonnet-20241022 \
  --api-key sk-ant-your-key \
  --max-tokens 4000 \
  --temperature 0.1

Use Cases:

Long document analysis
Code review and refactoring
Research and writing
Safety-critical applications

Google (Gemini)

Google’s Gemini models provide strong multimodal capabilities and competitive performance.

Available Models:

gemini-2.0-flash-exp - Latest experimental model
gemini-1.5-pro - High-capability model
gemini-1.5-flash - Fast and efficient

Configuration:

# Basic setup
rawi configure --provider google \
  --model gemini-2.0-flash-exp \
  --api-key your-google-api-key

# For multimodal tasks
rawi configure --provider google \
  --model gemini-1.5-pro \
  --api-key your-google-api-key \
  --temperature 0.4

Use Cases:

Multimodal tasks (text + images)
Search and information retrieval
Creative writing
Technical documentation

DeepSeek

Cost-effective AI with strong reasoning capabilities and competitive performance.

Available Models:

deepseek-chat - Versatile model for general use
deepseek-reasoner - Enhanced reasoning capabilities

Configuration:

# Basic setup
rawi configure --provider deepseek \
  --model deepseek-chat \
  --api-key sk-your-deepseek-key

# For reasoning tasks
rawi configure --provider deepseek \
  --model deepseek-reasoner \
  --api-key sk-your-deepseek-key \
  --temperature 0.2

Use Cases:

Cost-effective AI interactions
Complex reasoning tasks
Mathematical problem solving
Budget-conscious applications

Mistral

European AI platform with strong multilingual capabilities and GDPR compliance.

Available Models:

mistral-large-latest - Most capable model
mistral-small-latest - Efficient and fast
ministral-3b-latest - Lightweight model
ministral-8b-latest - Balanced performance
pixtral-large-latest - Multimodal capabilities

Configuration:

# Basic setup
rawi configure --provider mistral \
  --model mistral-large-latest \
  --api-key your-mistral-key

# For multilingual tasks
rawi configure --provider mistral \
  --model mistral-large-latest \
  --api-key your-mistral-key \
  --temperature 0.5

Use Cases:

European AI with GDPR compliance
Multilingual applications
JSON mode support
Function calling capabilities

Cerebras

Ultra-fast AI inference with Wafer-Scale Engine technology for exceptional speed.

Available Models:

llama3.1-70b - Meta’s Llama 3.1 70B model
llama3.1-8b - Meta’s Llama 3.1 8B model
llama-3.3-70b - Latest Meta Llama 3.3 70B

Configuration:

# Basic setup
rawi configure --provider cerebras \
  --model llama3.1-70b \
  --api-key csk-your-key

# For speed-critical tasks
rawi configure --provider cerebras \
  --model llama3.1-70b \
  --api-key csk-your-key \
  --temperature 0.7

Use Cases:

Ultra-fast inference requirements
High-throughput applications
Speed-critical workflows
Real-time AI interactions

Ollama (Local)

Run open-source models locally for privacy and control. No API key required.

Popular Models:

llama3.2:latest - Meta’s latest Llama model
mistral:latest - Mistral AI’s flagship model
codellama:latest - Specialized for code generation
phi3:latest - Microsoft’s compact model

Configuration:

# Install Ollama first: https://ollama.ai
curl -fsSL https://ollama.ai/install.sh | sh

# Start Ollama service
ollama serve

# Pull a model
ollama pull llama3.2

# Configure Rawi
rawi configure --provider ollama \
  --model llama3.2 \
  --base-url http://localhost:11434

Use Cases:

Privacy-sensitive tasks
Offline work environments
Cost-free AI interactions
Experimentation with open models

LM Studio (Local)

User-friendly local AI with GUI management for easy model handling and optimization.

Key Features:

🖥️ Easy GUI for model management
📦 Automatic model downloads
⚙️ Hardware optimization
🔧 Fine-tuning capabilities
📊 Performance monitoring

Popular Models:

Llama models (3.1, 3.2)
Mistral models
Code-specific models
Custom fine-tuned models

Configuration:

# Install LM Studio first: https://lmstudio.ai
# Download and install from website

# Download models through GUI
# Start local server in LM Studio

# Configure Rawi
rawi configure --provider lmstudio \
  --model your-loaded-model \
  --base-url http://localhost:1234

Setup Steps:

Download LM Studio from lmstudio.ai
Install and open the application
Browse and download models through the GUI
Start the local server
Configure Rawi to use LM Studio

Use Cases:

User-friendly local AI without command line
GUI-based model management
Hardware-optimized inference
Fine-tuning and customization
Educational and research purposes

Azure OpenAI

Enterprise-grade OpenAI models with enhanced security and compliance.

Available Models:

Same as OpenAI but with enterprise features
Custom fine-tuned models
Regional deployment options

Configuration:

# Basic setup
rawi configure --provider azure \
  --model gpt-4 \
  --api-key your-azure-key \
  --resource-name your-resource-name \
  --base-url https://your-resource.openai.azure.com

# With API version
rawi configure --provider azure \
  --model gpt-4 \
  --api-key your-azure-key \
  --resource-name your-resource-name \
  --api-version 2024-10-01-preview

Use Cases:

Enterprise environments
Compliance-critical applications
Custom model deployments
Regional data requirements

Amazon Bedrock

Access multiple AI providers through AWS’s managed service.

Available Models:

anthropic.claude-3-5-sonnet-20241022-v2:0 - Claude 3.5 Sonnet
anthropic.claude-3-haiku-20240307-v1:0 - Claude 3 Haiku
meta.llama3-2-90b-instruct-v1:0 - Llama 3.2 90B
amazon.titan-text-premier-v1:0 - Amazon Titan

Configuration:

# Using AWS credentials
export AWS_ACCESS_KEY_ID="your-access-key"
export AWS_SECRET_ACCESS_KEY="your-secret-key"
export AWS_REGION="us-east-1"

rawi configure --provider bedrock \
  --model anthropic.claude-3-5-sonnet-20241022-v2:0 \
  --region us-east-1

Use Cases:

AWS-integrated workflows
Multi-model experimentation
Enterprise AWS environments
Regulated industries

xAI (Grok)

Elon Musk’s xAI models with unique training and capabilities.

Available Models:

grok-2-1212 - Latest Grok model
grok-2-vision-1212 - Multimodal version
grok-beta - Beta testing model

Configuration:

rawi configure --provider xai \
  --model grok-2-1212 \
  --api-key xai-your-api-key

Use Cases:

Creative and experimental tasks
Alternative perspective generation
Research and exploration
Unique reasoning approaches

Provider Selection Guide

Recommended: OpenAI GPT-4o or Anthropic Claude 3.5 Sonnet

# OpenAI for general coding
rawi configure --provider openai --model gpt-4o

# Claude for code review
rawi configure --provider anthropic --model claude-3-5-sonnet-20241022

Recommended: Ollama with CodeLlama or Llama 3.2

# Install and setup
ollama pull codellama:13b
rawi configure --provider ollama --model codellama:13b

For Content Creation

Writing & Research
Multimodal Content

Recommended: Anthropic Claude 3.5 Sonnet for analysis, OpenAI GPT-4o for creativity

# Long-form analysis
rawi configure --profile research \
  --provider anthropic \
  --model claude-3-5-sonnet-20241022 \
  --max-tokens 4000

# Creative writing
rawi configure --profile creative \
  --provider openai \
  --model gpt-4o \
  --temperature 0.8

Recommended: Google Gemini or OpenAI GPT-4o

# Google Gemini for multimodal
rawi configure --provider google --model gemini-2.0-flash-exp

# OpenAI for image analysis
rawi configure --provider openai --model gpt-4o

For Enterprise Use

Security & Compliance
Privacy & Control

Recommended: Azure OpenAI or Amazon Bedrock

# Azure for Office 365 integration
rawi configure --provider azure \
  --model gpt-4 \
  --resource-name company-openai

# Bedrock for AWS environments
rawi configure --provider bedrock \
  --model anthropic.claude-3-5-sonnet-20241022-v2:0

Recommended: Ollama for complete data control

# Local deployment
ollama pull llama3.2:70b
rawi configure --provider ollama --model llama3.2:70b

Multi-Provider Workflows

You can use different providers for different tasks within the same project:

# Code generation with OpenAI
rawi configure --profile code \
  --provider openai \
  --model gpt-4o

# Code review with Claude
rawi configure --profile review \
  --provider anthropic \
  --model claude-3-5-sonnet-20241022

# Documentation with Gemini
rawi configure --profile docs \
  --provider google \
  --model gemini-1.5-pro

# Privacy-sensitive tasks with Ollama
rawi configure --profile private \
  --provider ollama \
  --model llama3.2

Getting API Keys

Choose your provider(s) based on your needs
Get API keys:
- OpenAI: platform.openai.com
- Anthropic: console.anthropic.com
- Google: aistudio.google.com
- xAI: console.x.ai
Set up enterprise providers:
- Azure: Contact your Azure administrator
- Bedrock: Set up through AWS Console
Install local providers:
- Ollama: ollama.ai

Provider-Specific Tips

OpenAI

Use gpt-4o for best overall performance
Set lower temperature (0.1-0.3) for factual tasks
Use o1-preview for complex reasoning problems

Anthropic

Claude excels at long document analysis
Use higher token limits (3000-4000) for detailed work
Great for safety-critical applications

Google

Gemini 2.0 Flash is experimental but very capable
Strong multimodal capabilities
Good for search and information tasks

Ollama

Start with llama3.2:8b for general use
Use codellama specifically for programming
Models run entirely on your machine

LM Studio

Use the GUI for easy model management
Great for users who prefer visual interfaces
Automatic hardware optimization
Easy model switching through interface

Azure/Bedrock

Check regional availability for models
Consider compliance requirements
May have different pricing structures

Troubleshooting Providers

Common Issues

API Key Errors:

# Test your configuration
rawi configure --test

# Verify API key format
rawi configure --show

Connection Problems:

# Check provider status
rawi info --providers

# Test with simple query
rawi ask "hello" --verbose

Model Not Available:

# List available models for provider
rawi info --models --provider openai

# Try different model
rawi configure --model gpt-3.5-turbo

Provider-Specific Troubleshooting

Ollama not responding:

# Check if Ollama is running
curl http://localhost:11434/api/tags

# Start Ollama service
ollama serve

# Pull model if missing
ollama pull llama3.2

LM Studio connection issues:

# Check if LM Studio server is running
curl http://localhost:1234/v1/models

# Start server in LM Studio app
# Check Models tab for loaded models
# Verify server is running on correct port

# Test connection
rawi configure --provider lmstudio --test

Azure authentication:

# Verify resource name and endpoint
rawi configure --provider azure --show

# Check Azure subscription
az account show

Next Steps

Configure your first provider: Configuration Guide
Set up environment variables: Environment Variables
Explore advanced settings: Advanced Options
Start using Rawi: Quick Start Guide