humanllm

OpenAI API-compatible LLM powered by YOU

Requirements

• Node.js

Usage

1. Register OpenAI API compatible endpoint http://localhost:3000/v1 to your chat app or agent

2. Run UI and API server

# UI: http://localhost:5173/
# API Server: http://localhost:3000/
$ npm run dev

3. Open UI http://localhost:5173/

4. Open the app or the agent and send a request

5. Read the request on the UI and respond to it!

Example: Call from Codex

# ~/.codex/humanllm.config.toml
model = "human"
model_provider = "humanllm"

[model_providers.humanllm]
name = "Human LLM"
base_url = "http://localhost:3000/v1"
bearer_token_env_var = "HUMANLLM_API_KEY"

[projects."/home/syuparn/tmp/hoge"]
trust_level = "trusted"

[tui.model_availability_nux]
"gpt-5.5" = 4

$ codex --profile humanllm