Coding assistants
You can use Privatemode with the AI coding assistants in your IDE such as Cline or Continue to provide secure and privacy-preserving agentic coding capabilities or tab completions.
- Continue
- Cline
Continue offers code completions, chat, and code editing capabilities. To set up Continue with Privatemode, first ensure the privatemode-proxy is set up as explained in the quickstart guide. Continue works best with prompt caching enabled.
Then, install the Continue extension for Visual Studio Code.
Once the extension is installed, edit ~/.continue/config.yaml
, and add the following entry to the models
key:
- name: Privatemode QwenCoder 14B
provider: vllm
model: qwen/qwen2.5-coder-14b-instruct-awq
apiKey: dummy
apiBase: http://localhost:8080/v1
defaultCompletionOptions:
maxTokens: 20
roles:
- autocomplete
- chat
- edit
- apply
You can find more information on Continue's configuration format in the official documentation.
Now you can select the "Privatemode QwenCoder 14B" model in the model selector of the extension, using Continue's code completions and its agentic chat with the Privatemode AI backend:

Cline is an advanced coding agent that can make complex code changes across files. Note that it's in an early stage and not always performs at its best. For best performance, make sure you are using the latest version!
To set up Cline with Privatemode, first ensure the privatemode-proxy is set up as explained in the quickstart guide. Cline works best with prompt caching enabled.
Then, install the Cline extension for Visual Studio Code.
Once the extension is installed, click "Select Model / API Provider" and change the configuration to the following:
- API Provider: OpenAI Compatible
- Base URL:
http://localhost:8080/v1
(Change the port number accordingly if you've set up the privatemode-proxy to use a different port) - OpenAI Compatible API Key:
null
(The privatemode-proxy takes care of authentication) - Model ID:
qwen/qwen2.5-coder-14b-instruct-awq
- Supports Images:
No
- Supports Browser Use:
No
- Context Window Size:
28000
Open the MODEL CONFIGURATION tab to see all settings. Note that the Qwen model supports a 32k context window but Cline may exceed that window such that you have to limit it to 28k context as described.
You can leave the other settings as-are:

You can find more information on Cline's model configuration in the official documentation
This allows you to use Cline's agentic capabilities with the Privatemode AI backend.