AI Providers Guide

Configure Anouk to work with your preferred AI provider.

Supported Providers

Anouk works with any OpenAI-compatible API. Here are the officially supported providers:

Provider	Local/Cloud	Free Tier	Notes
OpenAI	Cloud	No	Industry standard
Together.xyz	Cloud	Yes	Default provider
Anthropic	Cloud	No	Claude models
Ollama	Local	Yes	Privacy-focused
Hugging Face	Cloud	Yes	Open models

OpenAI

The industry-standard provider with GPT-4 and GPT-3.5 models.

Setup

Get an API key from platform.openai.com
Configure Anouk:

const aiService = new AIService({
  provider: 'openai',
  apiKey: 'sk-your-api-key',
  baseUrl: 'https://api.openai.com/v1',
  model: 'gpt-4'
});

Available Models

Model	Context	Best For
`gpt-4`	8K	Complex reasoning
`gpt-4-turbo`	128K	Long documents
`gpt-3.5-turbo`	16K	Fast, cost-effective

Pricing Note

OpenAI charges per token. GPT-4 is significantly more expensive than GPT-3.5-turbo.

Together.xyz

Default provider with competitive pricing and open-source models.

Setup

Create an account at together.xyz
Get your API key from the dashboard
Configure:

const aiService = new AIService({
  provider: 'together',
  apiKey: 'your-together-key',
  baseUrl: 'https://api.together.xyz/v1',
  model: 'meta-llama/Llama-3-70b-chat-hf'
});

Available Models

Model	Parameters	Notes
`meta-llama/Llama-3-70b-chat-hf`	70B	Recommended
`mistralai/Mixtral-8x7B-Instruct-v0.1`	8x7B	Fast
`togethercomputer/CodeLlama-34b-Instruct`	34B	Code-focused

Anthropic Claude

Anthropic's Claude models are known for safety and helpfulness.

Setup

Get API access at anthropic.com
Configure using the OpenAI-compatible endpoint:

const aiService = new AIService({
  provider: 'anthropic',
  apiKey: 'sk-ant-your-key',
  baseUrl: 'https://api.anthropic.com/v1',
  model: 'claude-3-opus-20240229'
});

Available Models

Model	Context	Best For
`claude-3-opus-20240229`	200K	Complex tasks
`claude-3-sonnet-20240229`	200K	Balanced
`claude-3-haiku-20240307`	200K	Fast, efficient

API Compatibility

Some Anthropic features may require additional headers. Check Anthropic's documentation for details.

Ollama (Local)

Run AI models locally for privacy and offline use.

Setup

Install Ollama from ollama.ai
Pull a model:

ollama pull llama2

Start the server:

ollama serve

Configure Anouk:

const aiService = new AIService({
  provider: 'ollama',
  apiKey: '',  // Not required for local
  baseUrl: 'http://localhost:11434/v1',
  model: 'llama2'
});

Available Models

# List available models
ollama list

# Popular models
ollama pull llama2
ollama pull mistral
ollama pull codellama
ollama pull phi

Benefits

Privacy: Data never leaves your machine
Free: No API costs
Offline: Works without internet
Fast: No network latency

Requirements

8GB+ RAM for 7B models
16GB+ RAM for 13B models
GPU recommended for larger models

Hugging Face

Access thousands of open-source models.

Setup

Create account at huggingface.co
Generate API token in settings
Configure:

const aiService = new AIService({
  provider: 'huggingface',
  apiKey: 'hf_your_token',
  baseUrl: 'https://api-inference.huggingface.co/v1',
  model: 'meta-llama/Llama-2-70b-chat-hf'
});

Pro Tier

Some models require a Pro subscription. Check model pages for access requirements.

Custom Providers

Connect to any OpenAI-compatible API:

const aiService = new AIService({
  provider: 'custom',
  apiKey: 'your-key',
  baseUrl: 'https://your-api.example.com/v1',
  model: 'your-model-name'
});

LM Studio

{
  provider: 'lmstudio',
  baseUrl: 'http://localhost:1234/v1',
  model: 'local-model'
}

Azure OpenAI

{
  provider: 'azure',
  baseUrl: 'https://your-resource.openai.azure.com',
  apiKey: 'your-azure-key',
  model: 'your-deployment-name'
}

Switching Providers

Runtime Switching

// Switch to a different provider
aiService.updateConfig({
  provider: 'openai',
  apiKey: 'sk-new-key',
  baseUrl: 'https://api.openai.com/v1',
  model: 'gpt-4'
});

Using Presets

import { ConfigManager } from 'anouk';

// Get preset for a provider
const openaiConfig = ConfigManager.getPresetConfigs()['openai'];

// Apply preset (remember to add your API key)
aiService.updateConfig({
  ...openaiConfig,
  apiKey: 'your-key'
});

Provider Comparison

Feature	OpenAI	Together	Ollama	Claude
Speed	Fast	Fast	Varies	Fast
Cost	$$$$	$$	Free	$$$
Privacy	Cloud	Cloud	Local	Cloud
Models	GPT	Open	Open	Claude
Context	128K	32K	32K	200K

Troubleshooting

Connection Refused (Ollama)

# Ensure Ollama is running
ollama serve

# Check if model is downloaded
ollama list

401 Unauthorized

Verify API key is correct
Check key hasn't expired
Ensure key has required permissions

429 Rate Limited

Reduce request frequency
Anouk includes built-in rate limiting (1s delay)
Consider upgrading your API plan

Model Not Found

Verify model name is exact
Check provider documentation for available models
Some models require special access