Does it support Qwen series hosted model?

[ruiqurm]

Is your feature request related to a problem? Please describe.

No response

Describe the solution you'd like

Hi. I want to use the Qwen-turbo model on the open interpreter. I checked the documentation and did not find the Qwen series hosted model. How can I add support for it? Is there a tutorial or guide to follow?
Thanks.

Describe alternatives you've considered

No response

Additional context

No response

[ebowwa]

Certainly! Let's dive straight into a concrete example of how you can integrate DeepSeek V3 from OpenRouter.ai into your Open Interpreter setup. This will help you leverage DeepSeek's powerful language model capabilities within Open Interpreter seamlessly.

Step-by-Step Integration Guide

1. Obtain Your OpenRouter.ai API Key

First, you need to sign up for an account on [OpenRouter.ai](https://openrouter.ai/) and generate an API key.

1. Sign Up / Log In:

   • Go to [OpenRouter.ai](https://openrouter.ai/) and create an account or log in if you already have one.

2. Generate API Key:

   • Navigate to the [API Keys](https://openrouter.ai/account/api-keys) section in your account dashboard.
   • Click on "Create New API Key".
   • Name your API key (e.g., OpenInterpreter-DeepSeek) and copy the generated key. Keep it secure!

2. Install Open Interpreter

If you haven't installed Open Interpreter yet, you can do so using pip:

pip install open-interpreter

3. Configure Open Interpreter to Use DeepSeek V3 via OpenRouter.ai

You have two primary ways to configure Open Interpreter to use DeepSeek V3: Command Line and Python Script.

---

A. Using Command Line

Open your terminal and run the following command, replacing YOUR_OPENROUTER_API_KEY with the API key you obtained:

interpreter --api_base "https://openrouter.ai/api/v1" --api_key "YOUR_OPENROUTER_API_KEY" --model "deepseek/deepseek-chat"

Example:

interpreter --api_base "https://openrouter.ai/api/v1" --api_key "sk-1234567890abcdef" --model "deepseek/deepseek-chat"

This command sets the API base to OpenRouter's endpoint, provides your API key, and specifies the DeepSeek V3 model.

---

B. Using a Python Script

If you prefer configuring within a Python environment, follow these steps:

1. Create a Python Script (e.g., use_deepseek.py):

from interpreter import interpreter

# Configure the interpreter to use DeepSeek V3 via OpenRouter.ai
interpreter.llm.model = "deepseek/deepseek-chat"
interpreter.llm.api_key = "YOUR_OPENROUTER_API_KEY"
interpreter.llm.api_base = "https://openrouter.ai/api/v1"

# Start an interactive chat session
interpreter.chat()

2. Replace the API Key:

Ensure you replace "YOUR_OPENROUTER_API_KEY" with your actual API key.

3. Run the Script:

python use_deepseek.py

This script initializes Open Interpreter with DeepSeek V3 and starts an interactive chat session.

---

4. Testing the Integration

After configuring, it's essential to verify that the integration works as expected.

1. Start Open Interpreter:

   • Command Line: If you used the command line method, running the command as shown above will start the interpreter.
   • Python Script: Running your Python script will start the interpreter.

2. Execute a Test Command:

   Once the interpreter is running, try a simple command to ensure everything is functioning.

   Example:

   interpreter> Plot the stock prices of AAPL and MSFT for the last month.

   Expected Outcome:

   The interpreter should process the command using DeepSeek V3 and generate the corresponding plot.

---

5. Advanced Configuration (Optional)

You can further customize your setup based on your requirements.

• Setting Environment Variables:

  Instead of passing the API key directly in commands or scripts, you can set it as an environment variable for better security.

  export OPENROUTER_API_KEY="YOUR_OPENROUTER_API_KEY"
  interpreter --api_base "https://openrouter.ai/api/v1" --api_key "$OPENROUTER_API_KEY" --model "deepseek/deepseek-chat"

• Using a .env File:

  Create a .env file in your project directory:

  OPENROUTER_API_KEY=your_api_key_here

  Then, modify your Python script to load environment variables:

  import os
  from interpreter import interpreter
  from dotenv import load_dotenv
  
  load_dotenv()  # Load variables from .env
  
  interpreter.llm.model = "deepseek/deepseek-chat"
  interpreter.llm.api_key = os.getenv("OPENROUTER_API_KEY")
  interpreter.llm.api_base = "https://openrouter.ai/api/v1"
  
  interpreter.chat()

  Install python-dotenv if not already installed:

  pip install python-dotenv

---

6. Troubleshooting Tips

• Invalid API Key:

  • Ensure that the API key is correctly copied without any extra spaces or characters.

• Network Issues:

  • Verify your internet connection.
  • Ensure that https://openrouter.ai/api/v1 is accessible from your network.

• Model Not Found:

  • Double-check the model name. It should be exactly "deepseek/deepseek-chat".

• Verbose Mode:

  • Enable verbose mode to get detailed logs which can help in debugging.

  Command Line:

  interpreter --api_base "https://openrouter.ai/api/v1" --api_key "YOUR_OPENROUTER_API_KEY" --model "deepseek/deepseek-chat" --verbose

  Python Script:

  interpreter.verbose = True

---

Conclusion

By following the above steps, you should have successfully integrated DeepSeek V3 via OpenRouter.ai into your Open Interpreter setup. This integration harnesses the advanced capabilities of DeepSeek, providing a more powerful and flexible language model experience.

If you encounter any further issues or have specific questions, feel free to reach out by commenting on [Issue #1572](#1572) or joining the [Open Interpreter Discord](https://discord.com/invite/open-interpreter) for real-time support.

Happy coding!

[tristayunsub]

wow so we can use o1, deepseek whatever we want! thank you

[blakkd]

Is your feature request related to a problem? Please describe.

No response

Describe the solution you'd like

Hi. I want to use the Qwen-turbo model on the open interpreter. I checked the documentation and did not find the Qwen series hosted model. How can I add support for it? Is there a tutorial or guide to follow? Thanks.

Describe alternatives you've considered

No response

Additional context

No response

Using the 1.0 preview, you can achieve what you look for using ollama: interpreter --model ollama/qwen2.5-coder:32b-instruct-q4_K_M.
No even need to pass --api-base or --provider litellm will figure out, syntax is that simple.

Just pay attention for now, only ollama/[modelname] is working, ollama_chat/[modelname] doesn't.

If anyone can help reporting the above chatgpt bullshit bot, so annoying...

[EDIT: my bad, I didn't see you were looking for the qwen API!]

[thedunston]

Also, openrouter isn't needed. DeepSeek has an OpenAI compatible AI:

interpreter --api_base "https://api.deepseek.com" --api_key "YOUR_KEY" --model "deepseek-chat"

[guxiangming]

Also, openrouter isn't needed. DeepSeek has an OpenAI compatible AI:

interpreter --api_base "https://api.deepseek.com" --api_key "YOUR_KEY" --model "deepseek-chat"

How to persist Settings? You do not need to declare every time you start