Enter your endpoint and API key in Settings, then start chatting.
Delete conversation?
This conversation will be permanently removed from this browser.
Privacy & Data
Before using SBCODE vLLM Client, please understand where your information is stored and where your chat requests are sent.
This client runs in your browser
SBCODE vLLM Client is a client-side application. The settings and conversations used by this client are stored locally in your web
browser using localStorage.
Information stored locally
Your vLLM endpoint address
Your Bearer API key
Your selected model
Your saved conversations and chat history
SBCODE does not receive your chat data
The client does not send your API key, saved conversations, prompts, or vLLM responses to the SBCODE web server as part of normal
operation. After this webpage has loaded, chat requests are sent directly from your browser to the vLLM endpoint you configure.
RunPod / vLLM
When you send a message, the message and the conversation history included in that request are sent to the RunPod/vLLM endpoint you
configured. That service and the infrastructure running your endpoint may process, monitor, or retain information according to its own
policies, configuration, and logging systems.
Important:
Do not enter passwords, private keys, personal information, confidential business information, or other sensitive data unless you understand
and accept the risks of sending that information to the RunPod/vLLM service you have configured.
Website and Cloudflare
Your browser must connect to the SBCODE website to download this client. The website may use Cloudflare and normal web-server
infrastructure, which can process standard website/network information associated with those requests. This is separate from the client
sending your chat requests directly to your configured vLLM endpoint.
Clearing your local data
The client uses your browser's localStorage. Clearing this site's stored browser data will remove the locally saved settings and
conversations. This does not delete information that may already have been sent to a RunPod/vLLM endpoint.