DEDUPED Hot AI repos. Real use cases. What people build. Get the Monday 5

How to point LLM at any OpenAI-compatible endpoint with LLM

Use uvx to run prompts or chats against LM Studio or another compatible server with no setup.

Short answer: Start your OpenAI-compatible server, for example LM Studio; then run uvx llm openai endpoint http://localhost:1234/v1 with -m and your model name followed by your prompt; add --chat for an interactive session. All steps.

Who it is forA developer trying out models served by LM Studio or a similar local server.

The problemConfiguring each server as a model takes time.

The outcomeA prompt or chat answered by your endpoint without installing LLM first.

What you need first

  • Python with pip, or Homebrew, pipx or uv
  • An API key for a hosted model, or a local model server such as Ollama or LM Studio
  • Plugins for providers such as Gemini

Steps

  1. Start your OpenAI-compatible server, for example LM Studio
  2. Run uvx llm openai endpoint http://localhost:1234/v1 with -m and your model name followed by your prompt
  3. Add --chat for an interactive session
  4. Add --key your-api-key if the endpoint needs authentication

Prompt for your agent

Using simonw/llm via uvx, run a prompt against my endpoint <url> and model <name> with `uvx llm openai endpoint`, then start a `--chat` session.

Caveat. Exact multi-line command is in the README quick start.

Use this in your AI agent

Ask your agent

Using the Deduped MCP, get the setup steps for simonw/llm and "Point LLM at any OpenAI-compatible endpoint"

Add Deduped to Claude Code once

claude mcp add --transport http deduped https://dedupedai.com/mcp

Cursor, ChatGPT, others

Share: X LinkedIn Hacker News Email Copy link

Third-party repos we don't own. AI-written use cases can be wrong. Disclaimer