DOCUMENTATION / QUICK START
Start in your terminal.
Endpoint Doctor is a local-first diagnostic CLI for OpenAI-compatible LLM APIs.
1. Install
pip install apiwellsUse Python 3.10 or newer with pip. Verify the installation with apiwells --version.
2. Run locally
Set APIWELLS_API_KEY locally using your shell's secret-input mechanism, then supply your exact API base URL:
apiwells doctor --base-url https://YOUR-API-HOST/v1Replace the placeholder with your endpoint. Include the exact path prefix; /v1 is not added automatically. Basic mode checks connectivity, authentication and the model list without inference requests.
To check chat, streaming, usage, tool calling and structured output, explicitly select Deep mode and a model:
apiwells doctor --base-url https://YOUR-API-HOST/v1 --deep --model YOUR-MODEL-IDReplace YOUR-MODEL-ID with an available model. Deep mode sends multiple potentially billable inference requests. Add --v2-json for a structured report.
Keep API keys in your local environment. Do not submit keys, tokens, or sensitive response data to this website or in public feedback.
3. Read the report
Review the result of each check to understand your endpoint's behavior. The landing page shows a recorded run from Endpoint Doctor 0.2.1 with qwen3.8-flash, supplied by the project owner. It is not a live test; results and timings are specific to that run. Host and IP details are omitted.
View Example Report →