OpenAI-Compatible APIs

Connect Page Assist to OpenAI-compatible API providers

Page Assist supports any service that implements the OpenAI API specification. This includes cloud providers, self-hosted services, and other local inference servers.

Supported Providers

Page Assist includes built-in support for many popular providers:

Cloud Providers

• OpenAI - GPT models (GPT-4, GPT-3.5, etc.)
• Anthropic (Claude) - Claude models
• Google AI - Gemini models
• Groq - Fast inference API
• DeepSeek - DeepSeek models
• Fireworks - Fast inference platform
• Together AI - Open source models
• OpenRouter - Unified API gateway
• Mistral - Mistral AI models
• xAI - Grok models
• Novita AI - AI inference platform
• Hugging Face - Inference API
• Infinigence AI - AI platform
• SiliconFlow - AI inference
• Moonshot - AI platform
• BigModel (Zhipu) - Chinese AI models
• MiniMax - Chinese AI models
• VolcEngine - ByteDance AI
• TencentCloud - Tencent AI
• AlibabaCloud - Alibaba AI

Self-Hosted / Local

• LLaMA.cpp - C++ implementation
• Llamafile - Single-file executables
• vLLM - High-performance inference
• Ollama (secondary instances)
• Vercel AI Gateway - AI proxy service
• Chutes - Custom deployments

Adding a Provider

1. Open Settings

   Click the Page Assist icon in your browser toolbar, then click the Settings icon.

2. Navigate to OpenAI Compatible API

   Go to the “OpenAI Compatible API” tab.

3. Click Add Provider

   Click the “Add Provider” button.

4. Select Provider

   Choose your provider from the dropdown:

   • If listed, select it directly (e.g., “OpenAI”, “Anthropic”, “Groq”)
   • If not listed, select “Custom”
5. Enter Configuration

   Provide the required information:

   OpenAI

   Base URL: https://api.openai.com/v1
   API Key: sk-...
   

   Base URL: https://api.anthropic.com/v1
   API Key: sk-ant-...
   

   Base URL: https://api.groq.com/openai/v1
   API Key: gsk_...
   

   Base URL: https://your-api.example.com/v1
   API Key: your-api-key (if required)
   
6. Add Models (if required)

   Some providers require manual model entry:

   • Auto-detected: Ollama, LM Studio, LLaMA.cpp, Llamafile, vLLM
   • Manual entry: Most cloud providers

   For manual entry, click “Add Model” and enter model names like:

   • gpt-4-turbo
   • claude-3-5-sonnet-20241022
   • llama-3.1-70b-versatile
7. Save Configuration

   Click “Save” to store your provider configuration.

Provider-Specific Setup

OpenAI

1. Get API Key

   1. Visit platform.openai.com
   2. Sign in or create an account
   3. Navigate to API Keys
   4. Create a new API key
2. Configure in Page Assist

   • Base URL: https://api.openai.com/v1
   • API Key: Your OpenAI API key (starts with sk-)
   • Models: gpt-4-turbo, gpt-4o, gpt-3.5-turbo, etc.

Anthropic (Claude)

1. Get API Key

   1. Visit console.anthropic.com
   2. Sign in or create an account
   3. Navigate to API Keys
   4. Create a new API key
2. Configure in Page Assist

   • Base URL: https://api.anthropic.com/v1
   • API Key: Your Anthropic API key (starts with sk-ant-)
   • Models: claude-3-5-sonnet-20241022, claude-3-opus-20240229, etc.

Groq

1. Get API Key

   1. Visit console.groq.com
   2. Sign in with your account
   3. Navigate to API Keys
   4. Create a new API key
2. Configure in Page Assist

   • Base URL: https://api.groq.com/openai/v1
   • API Key: Your Groq API key
   • Models: llama-3.1-70b-versatile, mixtral-8x7b-32768, etc.

LLaMA.cpp Server

1. Start LLaMA.cpp Server

   Run the LLaMA.cpp server with OpenAI API compatibility:

   ./server -m model.gguf --port 8080
   
2. Configure in Page Assist

   • Provider: Select “LLaMA.cpp”
   • Base URL: http://localhost:8080/v1 (or your custom URL)
   • API Key: Not required for local instances
   • Models: Auto-detected from running server

vLLM

1. Start vLLM Server

   Launch vLLM with OpenAI-compatible API:

   python -m vllm.entrypoints.openai.api_server \
     --model meta-llama/Llama-2-7b-hf \
     --port 8000
   
2. Configure in Page Assist

   • Provider: Select “vLLM”
   • Base URL: http://localhost:8000/v1
   • API Key: Not required for local instances
   • Models: Auto-detected

Custom Provider

For any unlisted OpenAI-compatible service:

1. Select Custom

   Choose “Custom” from the provider dropdown.

2. Enter Details

   • Provider Name: Give it a recognizable name
   • Base URL: Your API endpoint (must end with /v1)
   • API Key: If required by your service
3. Add Models

   Manually add model names or let Page Assist attempt auto-detection.

Model Management

Auto-Detection

The following providers support automatic model detection:

• Ollama (secondary instances)
• LM Studio
• LLaMA.cpp
• Llamafile
• vLLM

Page Assist will automatically fetch available models when you save the provider.

Manual Model Entry

For most cloud providers, you need to manually add models:

1. After saving the provider, click “Add Model”
2. Enter the exact model identifier
3. Optionally specify if it’s a chat or embedding model
4. Save the model

Model Types

Specify the model type for proper functionality:

• Chat Models: For conversations (GPT-4, Claude, etc.)
• Embedding Models: For RAG features (text-embedding-ada-002, etc.)

Authentication

API Keys

Most cloud providers require API keys:

OpenAI: sk-...
Anthropic: sk-ant-...
Groq: gsk_...

API keys are stored securely in your browser’s extension storage.

Custom Headers

Some providers require custom headers:

1. In provider settings, expand “Advanced Options”
2. Add custom headers as key-value pairs:
   Key: X-Custom-Header
   Value: your-value
   

Bearer Tokens

For services using bearer tokens:

1. Enter the token in the “API Key” field
2. Or add a custom header: Authorization: Bearer your-token

Troubleshooting

Connection Failed

1. Verify Base URL

   Ensure the URL is correct and includes /v1 at the end (if required).

2. Check API Key

   Verify your API key is valid and has sufficient credits/permissions.

3. Test Endpoint

   Use curl or Postman to test the endpoint directly:

   curl https://api.openai.com/v1/models \
     -H "Authorization: Bearer YOUR_API_KEY"
   
4. Check CORS

   Some self-hosted services may need CORS configuration to work with browser extensions.

Models Not Loading

1. Check if models need manual entry for your provider
2. Verify the API endpoint supports /v1/models for auto-detection
3. Try manually adding model names
4. Check API key permissions

Authentication Errors

1. Verify API key format is correct
2. Check for extra spaces in the API key
3. Ensure the API key has necessary permissions
4. Verify the API key hasn’t expired

Rate Limiting

If you encounter rate limit errors:

1. Wait before sending more requests
2. Check your provider’s rate limits
3. Consider upgrading your plan
4. Use a different provider for high-volume usage

Best Practices

1. Secure Your API Keys: Never share your API keys publicly
2. Monitor Usage: Track API usage to avoid unexpected charges
3. Test with Small Models: Start with smaller/cheaper models for testing
4. Set Usage Limits: Configure spending limits in your provider’s dashboard
5. Use Multiple Providers: Distribute usage across providers for reliability
6. Keep Keys Rotated: Regularly rotate API keys for security

Provider Comparison

Provider	Strengths	Best For
OpenAI	Most capable models	Complex tasks, reasoning
Anthropic	Long context, safety	Document analysis, safe AI
Groq	Ultra-fast inference	Speed-critical applications
Local (Ollama/LM Studio)	Privacy, no costs	Private data, development
Together AI	Open models, good pricing	Open source model access
OpenRouter	Access to many models	Model variety, fallbacks

Next Steps

• Learn about Ollama setup
• Explore LM Studio integration
• Set up Chrome AI
• Configure Knowledge Base features