Skip to content
This repository was archived by the owner on Apr 28, 2026. It is now read-only.

provider.md

Latest commit

 

History

History
301 lines (212 loc) · 6.71 KB

provider.md

File metadata and controls

301 lines (212 loc) · 6.71 KB

Provider Command

Explore and discover AI providers, their models, and capabilities supported by Rawi.

Navigation

Basic Usage

# List all supported providers
rawi provider --list

# Show models for a specific provider
rawi provider --list-models openai

# Interactive provider explorer
rawi provider

Command Options

Option Description
-l, --list Display all supported AI providers
-m, --list-models <provider> Show available models for a specific provider
--help Show command help and examples

Examples

1. List All Providers

Get an overview of all supported AI providers:

rawi provider --list

Shows:

  • Provider names and icons
  • Current status (configured/available)
  • Key features and capabilities
  • Setup requirements

2. Explore Provider Models

View available models for a specific provider:

# OpenAI models
rawi provider --list-models openai

# Anthropic models
rawi provider --list-models anthropic

# Local Ollama models
rawi provider --list-models ollama

3. Interactive Mode

Run without options for an interactive explorer:

rawi provider

This opens an interactive menu to:

  • Browse providers
  • View model details
  • Compare capabilities
  • Get setup instructions

Supported Providers

Cloud Providers

Provider Icon Models Key Features
OpenAI 🔵 GPT-4, GPT-3.5 Industry leading, versatile
Anthropic 🟣 Claude-3, Claude-2 Safety-focused, long context
Google 🔴 Gemini Pro, PaLM Multimodal, code-optimized
Azure OpenAI 🔷 GPT-4, GPT-3.5 Enterprise, compliance
Amazon Bedrock 🟠 Multiple models AWS integration, variety
xAI ⚫ Grok models Real-time, conversational

Local Providers

Provider Icon Models Key Features
Ollama 🟢 Llama2, Mistral, CodeLlama Privacy, offline, fast
LM Studio 🔗 Various local models GUI, model management

Provider Details

OpenAI

🔵 OpenAI
├─ Models: gpt-4, gpt-4-turbo, gpt-3.5-turbo
├─ Context: Up to 128k tokens
├─ Features: Function calling, vision, code interpreter
├─ Setup: API key required
└─ Cost: Pay-per-use

Anthropic

🟣 Anthropic
├─ Models: claude-3-opus, claude-3-sonnet, claude-3-haiku
├─ Context: Up to 200k tokens
├─ Features: Constitutional AI, safety-focused
├─ Setup: API key required
└─ Cost: Pay-per-use

Ollama (Local)

🟢 Ollama
├─ Models: llama2, mistral, codellama, phi, gemma
├─ Context: Varies by model
├─ Features: Completely offline, privacy-first
├─ Setup: Local installation required
└─ Cost: Free (compute only)

Model Information

When viewing models for a provider:

rawi provider --list-models openai

Output includes:

  • Model name and version
  • Context window size
  • Capabilities (text, code, vision)
  • Recommended use cases
  • Performance characteristics
  • Cost information

Setup Requirements

API-Based Providers

Most cloud providers require:

  1. Account registration
  2. API key generation
  3. Configuration in Rawi
# Configure after getting API key
rawi configure --profile myprofile

Local Providers

For Ollama and LM Studio:

  1. Install the software locally
  2. Download/pull models
  3. Ensure service is running
# Example Ollama setup
ollama pull llama2
ollama serve

# Then configure in Rawi
rawi configure --profile local

Use Cases

1. Provider Research

Before configuring, explore what's available:

rawi provider --list
rawi provider --list-models anthropic

2. Model Comparison

Compare models across providers:

# Check OpenAI options
rawi provider --list-models openai

# Compare with Anthropic
rawi provider --list-models anthropic

# Consider local alternatives
rawi provider --list-models ollama

3. Capability Discovery

Find providers with specific features:

  • Vision capabilities
  • Long context windows
  • Code generation
  • Multilingual support

Integration Examples

Scripting with Provider Info

#!/bin/bash
# Check if local provider is available
if rawi provider --list-models ollama | grep -q "llama2"; then
    echo "Using local Ollama"
    rawi ask --profile local "Hello world"
else
    echo "Falling back to OpenAI"
    rawi ask --profile cloud "Hello world"
fi

Configuration Automation

# Auto-configure based on available providers
for provider in openai anthropic ollama; do
    if rawi provider --list-models $provider >/dev/null 2>&1; then
        echo "$provider is available"
        # Configure if desired
    fi
done

Related Commands

  • configure — Set up providers after exploration
  • info — See currently configured providers
  • ask — Use specific providers for conversations

Troubleshooting

Provider Not Available

If a provider doesn't appear in the list:

  1. Check your internet connection (for cloud providers)
  2. Verify local installations (Ollama, LM Studio)
  3. Update Rawi to the latest version

No Models Listed

When --list-models returns empty:

  • Cloud providers: Check API key configuration
  • Ollama: Ensure models are pulled and service is running
  • LM Studio: Verify models are downloaded and server is active

Connection Issues

For cloud providers:

# Test connectivity
curl -I https://api.openai.com/v1/models

# Check configuration
rawi info --profiles

For local providers:

# Test Ollama
curl http://localhost:11434/api/tags

# Check if service is running
ps aux | grep ollama

Next Steps: