Aider Integration Guide for Freeway

Freeway exposes an OpenAI-compatible local endpoint at http://localhost:8787/v1, so Aider works using standard OpenAI-compatible configuration. Depending on your setup, Aider may use OPENAI_API_BASE / --openai-api-base, while some other OpenAI-compatible tools may instead reference OPENAI_BASE_URL.

📋 Quick Start

# 1. Install Aider
pip install aider-chat

# 2. Set your base URL
export OPENAI_API_BASE="http://localhost:8787/v1"

# 3. Run Aider
aider --openai-api-base http://localhost:8787/v1 --openai-api-key <your_FREEWAY_API_KEY> --model llama-3.3-70b

Under 2 minutes. Done.

---

Why Use Aider with Freeway?

• Unified local endpoint - Single interface for multiple providers
• Provider routing - Smart request distribution
• Fallbacks - Automatic provider switching on failure

---

Common Endpoint

http://localhost:8787/v1

Prerequisites

• Freeway installed and running locally
• Python installed
• Aider installed

pip install aider-chat

Step 1: API Key Configuration

If FREEWAY_API_KEY authentication is enabled, use your configured key. If authentication is disabled, any non-empty placeholder value is typically sufficient.

export OPENAI_API_KEY="<your_FREEWAY_API_KEY>"

Step 2: Connect Aider to Freeway

Method 1: Command Line (Recommended)

aider --openai-api-base http://localhost:8787/v1 \
      --openai-api-key <your_FREEWAY_API_KEY> \
      --model llama-3.3-70b

Method 2: Environment Variables

Linux / macOS

export OPENAI_API_BASE="http://localhost:8787/v1"
export OPENAI_API_KEY="<your_FREEWAY_API_KEY>"
export AIDER_MODEL="llama-3.3-70b"

aider

Windows (Command Prompt)

set OPENAI_API_BASE=http://localhost:8787/v1
set OPENAI_API_KEY=<your_FREEWAY_API_KEY>
set AIDER_MODEL=llama-3.3-70b

aider

Windows (PowerShell)

$env:OPENAI_API_BASE="http://localhost:8787/v1"
$env:OPENAI_API_KEY="<your_FREEWAY_API_KEY>"
$env:AIDER_MODEL="llama-3.3-70b"

aider

Step 3: Model Selection

Note: The model IDs below are examples only. Use the model IDs returned by your local Freeway /v1/models endpoint or shown in the Freeway console for the most accurate list.

Use any Freeway-supported model ID with the --model flag:

# Groq models
--model groq/llama-3.3-70b
--model groq/llama-3.1-8b

# Freeway-supported models
--model llama-3.3-70b
--model llama-3.2-3b
--model mistral-7b

Step 4: Verification

curl http://localhost:8787/v1/models

Expected output: A JSON list of available models from your local Freeway instance.

Step 5: Test with a Real File

echo "print('Hello from Aider + Freeway')" > test.py

aider --openai-api-base http://localhost:8787/v1 \
      --openai-api-key <your_FREEWAY_API_KEY> \
      --model llama-3.3-70b \
      test.py

✅ Verification Checklist

• Aider runs without connection errors
• Aider shows the loaded file
• Aider responds to your requests

🔧 Troubleshooting

Problem	Solution
Connection refused	Ensure Freeway is running on port 8787
Invalid API key	Check your Freeway config
Model not found	Use Freeway-supported model ID
Aider: command not found	pip install aider-chat

📚 Next Steps

• Try different Freeway-supported models
• Configure routing and fallbacks in Freeway

🆘 Support

• Freeway Issues: Freeway documentation
• Aider Issues: Aider official documentation