Using Commercial APIs with Audify

Audify now supports commercial LLM APIs in addition to local Ollama models. This allows you to use powerful cloud-based models like DeepSeek, Claude, GPT-4, and Gemini for generating audiobook scripts.

Supported APIs

  • DeepSeek - Fast and cost-effective API

  • Anthropic Claude - High-quality reasoning and long context

  • OpenAI GPT - Industry-standard models

  • Google Gemini - Google’s latest AI models

Setup

1. Create a .keys File

Create a .keys file in the project root directory with your API keys:

cp .keys.example .keys

Edit the .keys file and add your API keys:

DEEPSEEK=sk-your-deepseek-api-key-here
ANTHROPIC=sk-ant-your-anthropic-api-key-here
OPENAI=sk-your-openai-api-key-here
GEMINI=your-google-api-key-here
# Or use GOOGLE (both names are equivalent):
# GOOGLE=your-google-api-key-here

Note: For Google Gemini, you can use either GEMINI or GOOGLE as the key name in your .keys file. Both are supported and equivalent. Similarly, when setting environment variables, both GEMINI_API_KEY and GOOGLE_API_KEY will work.

Important Security Notes:

  • The .keys file is already in .gitignore to prevent accidental commits

  • Never commit API keys to version control

  • Keep your API keys secure and private

  • You can also set API keys as environment variables (e.g., DEEPSEEK_API_KEY)

2. Get API Keys

DeepSeek

  1. Visit https://platform.deepseek.com/

  2. Sign up for an account

  3. Generate an API key from your dashboard

  4. DeepSeek offers competitive pricing and good performance

Anthropic Claude

  1. Visit https://console.anthropic.com/

  2. Create an account

  3. Generate an API key

  4. Claude excels at reasoning and has large context windows

OpenAI

  1. Visit https://platform.openai.com/

  2. Sign up and add billing information

  3. Generate an API key from the API section

  4. Models include GPT-4, GPT-4-turbo, GPT-3.5-turbo

Google Gemini

  1. Visit https://ai.google.dev/

  2. Get started with Gemini API

  3. Create an API key

  4. Gemini offers various models including Gemini Pro

Usage

To use a commercial API, prefix the model name with api: when using the -m or --llm-model option:

DeepSeek Examples

# Using DeepSeek Chat
python -m audify.create_audiobook mybook.epub -m "api:deepseek/deepseek-chat"

# Using DeepSeek R1 (reasoning model)
python -m audify.create_audiobook mybook.epub -m "api:deepseek/deepseek-reasoner"

Claude Examples

# Using Claude 3 Sonnet
python -m audify.create_audiobook mybook.epub -m "api:anthropic/claude-3-sonnet-20240229"

# Using Claude 3.5 Sonnet (latest)
python -m audify.create_audiobook mybook.epub -m "api:anthropic/claude-3-5-sonnet-20240620"

# Using Claude 3 Opus (most capable)
python -m audify.create_audiobook mybook.epub -m "api:anthropic/claude-3-opus-20240229"

OpenAI Examples

# Using GPT-4
python -m audify.create_audiobook mybook.epub -m "api:openai/gpt-4"

# Using GPT-4 Turbo
python -m audify.create_audiobook mybook.epub -m "api:openai/gpt-4-turbo-preview"

# Using GPT-3.5 Turbo (faster, cheaper)
python -m audify.create_audiobook mybook.epub -m "api:openai/gpt-3.5-turbo"

Google Gemini Examples

# Using Gemini Pro
python -m audify.create_audiobook mybook.epub -m "api:gemini/gemini-pro"

# Using Gemini 1.5 Pro
python -m audify.create_audiobook mybook.epub -m "api:gemini/gemini-1.5-pro"

Complete Example

# Create audiobook using DeepSeek with Spanish translation
python -m audify.create_audiobook mybook.epub \
  -m "api:deepseek/deepseek-chat" \
  -l en \
  -t es \
  -v af_bella \
  --save-scripts

# Create audiobook from PDF using Claude
python -m audify.create_audiobook document.pdf \
  -m "api:anthropic/claude-3-sonnet-20240229" \
  -l en \
  -v af_sarah

# Process directory of books using GPT-4
python -m audify.create_audiobook ./books/ \
  -m "api:openai/gpt-4-turbo-preview" \
  -l en \
  -o ./output

Comparing Ollama vs Commercial APIs

Feature

Ollama (Local)

Commercial APIs

Cost

Free

Pay per token

Privacy

Complete privacy

Data sent to cloud

Speed

Depends on hardware

Generally faster

Quality

Varies by model

State-of-the-art

Internet

Not required

Required

Setup

Install Ollama + models

Just API key

Model Recommendations

For Best Quality

  • Claude 3 Opus: Best reasoning and quality

  • GPT-4: Reliable and well-tested

  • Claude 3.5 Sonnet: Great balance of quality and speed

For Best Value

  • DeepSeek Chat: Very cost-effective

  • GPT-3.5 Turbo: Fast and affordable

  • Gemini Pro: Good balance

For Reasoning Tasks

  • DeepSeek R1: Specialized reasoning model

  • Claude 3 Opus: Excellent reasoning

  • GPT-4: Strong reasoning capabilities

Troubleshooting

API Key Not Found

Error: API key issue. Please ensure your API key is properly configured

Solution: Check that:

  1. The .keys file exists in the project root

  2. The API key is properly formatted (e.g., DEEPSEEK=sk-...)

  3. No extra spaces around the = sign

  4. The key name matches the service (DEEPSEEK, ANTHROPIC, OPENAI, GEMINI)

Connection Error

Error: Could not connect to commercial API

Solution:

  1. Check your internet connection

  2. Verify the API key is valid

  3. Check if the API service is experiencing downtime

  4. Ensure you have API credits/billing enabled

Invalid Model Name

Error: Failed to generate audiobook script

Solution:

  1. Verify the model name is correct (check API documentation)

  2. Ensure you’re using the api: prefix

  3. Some models may require specific API access tiers

Environment Variables (Alternative to .keys file)

Instead of using a .keys file, you can set environment variables:

# Bash/Zsh
export DEEPSEEK_API_KEY="sk-your-key"
export ANTHROPIC_API_KEY="sk-ant-your-key"
export OPENAI_API_KEY="sk-your-key"
# For Google Gemini, use either GOOGLE_API_KEY or GEMINI_API_KEY (both work):
export GOOGLE_API_KEY="your-key"
# export GEMINI_API_KEY="your-key"  # Alternative, equivalent to above

# Then run normally
python -m audify.create_audiobook mybook.epub -m "api:deepseek/deepseek-chat"

Cost Considerations

Commercial APIs charge based on token usage (input + output). Audiobook generation can use significant tokens due to:

  • Processing entire chapters

  • Generating narrative scripts

  • Multiple chapters per book

Tips to minimize costs:

  1. Use more cost-effective models like DeepSeek or GPT-3.5

  2. Limit --max-chapters for testing

  3. Monitor your API usage on the provider’s dashboard

  4. Consider using Ollama for testing, then commercial APIs for final output

Technical Details

The implementation uses LiteLLM to provide a unified interface across different API providers. This means:

  • Consistent API regardless of provider

  • Easy to add new providers

  • Automatic retry and error handling

  • Unified token counting

Model names are passed directly to LiteLLM, so any model supported by LiteLLM can be used with the api: prefix.