API Keys Setup Guide

This guide provides detailed instructions on how to obtain API keys for all LLM providers and tools used in AgentFlow.

---

1. OpenAI API Key

Purpose: Access OpenAI's language models (GPT-5, GPT-4o, etc.), used in AgentFlow for judging answer correctness.

How to obtain:

1. Visit OpenAI Platform
2. Sign up/Log in to your account
3. Navigate to API Keys page
4. Click "Create new secret key" to generate a new API key

Available models: OpenAI Models Documentation

Common model names: gpt-4o,gpt-4o-mini, gpt-4-turbo, gpt-3.5-turbo

---

2. Google API Key

Purpose: used in AgentFlow for Google Search tool.

How to obtain:

1. Visit Google AI Studio
2. Log in with your Google account
3. Click "Get API key"
4. Create or select a Google Cloud project
5. Copy the generated API key

---

3. DashScope API Key (Alibaba Cloud)

Purpose: Access Alibaba Cloud's Qwen (Tongyi Qianwen) model series. In AgentFlow, we use DashScope to call Qwen-2.5-7B-Instruct as the LLM engine for agents (except planner) and tools.

How to obtain:

1. Visit Alibaba Cloud DashScope Console
2. Log in with your Alibaba Cloud account
3. Navigate to API-KEY Management
4. Create a new API key

Official guide: Get API Key (you may need to translate the page to English using your browser's translation feature)

Available models: DashScope Model Documentation

Common model names: qwen-turbo, qwen-plus, qwen-max, qwen2.5-7b-instruct, qwen2.5-72b-instruct

Note: For international users, we recommend using Together AI to access Qwen-2.5-7B-Instruct model. Alternatively, you can serve the model locally using vLLM.

---

4. Together API Key - Recommended for More International Users

Purpose: Access open-source models on TogetherAI platform, including Qwen, Llama, Mixtral, etc.

Recommended for: International users who want to access Qwen-2.5-7B-Instruct and other open-source models with better global network connectivity.

How to obtain:

1. Visit Together.ai
2. Sign up/Log in to your account
3. Navigate to Settings > API Keys
4. Create a new API key

Available models: Together Models Documentation

Common model names:

• Qwen models: Qwen/Qwen2.5-7B-Instruct, Qwen/Qwen2.5-72B-Instruct
• Other models: meta-llama/Llama-3-70b-chat-hf, mistralai/Mixtral-8x7B-Instruct-v0.1

Important Note for Qwen Models: Together AI offers both Turbo (quantized) and standard (non-quantized) versions of Qwen models. For best performance and accuracy, we recommend using the non-quantized versions (e.g., Qwen/Qwen2.5-7B-Instruct instead of Qwen/Qwen2.5-7B-Instruct-Turbo). The Turbo versions are faster but may have reduced quality due to quantization.

Important Notes

1. Security: All API keys are sensitive information - never expose them or commit to public repositories
2. Costs: Most API services are paid - understand pricing before use
3. Quotas: Some services have free tiers or rate limits - monitor your usage
4. Environment Variables: Copy .env.template to .env and fill in your actual API keys
5. Regional Recommendations:
   • China users: Use DashScope for Qwen models
   • International users: Use Together AI for Qwen models

---

Quick Reference Table

API Key	Purpose	Sign Up Link	Documentation
OPENAI_API_KEY	OpenAI GPT models	platform.openai.com	Docs
GOOGLE_API_KEY	Gemini models	aistudio.google.com	Docs
DASHSCOPE_API_KEY	Qwen models	dashscope.console.aliyun.com	Docs
TOGETHER_API_KEY	Qwen & open-source models	together.ai	Docs