From 5972527e4591ebe9d7d6199f50fde61b000749b7 Mon Sep 17 00:00:00 2001 From: PerryLink <255665900+PerryLink@users.noreply.github.com> Date: Sun, 13 Sep 2026 17:03:36 +0800 Subject: [PATCH] docs(quickstart): add generic OpenAI-compatible endpoint section --- QUICKSTART.md | 22 ++++++++++++++++++++++ docs/configuration.md | 4 ++++ 2 files changed, 26 insertions(+) diff --git a/QUICKSTART.md b/QUICKSTART.md index 52e5216e9..2d35899ee 100644 --- a/QUICKSTART.md +++ b/QUICKSTART.md @@ -257,6 +257,28 @@ degrading to a different search method. You can replace OpenRouter with another OpenAI-compatible LLM endpoint by changing the `[llm]` model, base URL, and key. +## Use another OpenAI-compatible endpoint + +The walkthrough above uses OpenRouter, but the same `[llm]` settings work +with any OpenAI-compatible endpoint. For example, to use the official +DeepSeek API, set these three values in `~/.everos/everos.toml`: + +```toml +[llm] +model = "deepseek-flash" +api_key = "" +base_url = "https://api.deepseek.com" +``` + +Then start the server with `everos server start`. This is a config-only +change: no EverOS code changes are required. To switch models, update +`model` to another identifier offered by your endpoint and restart the +server. + +The same pattern applies to any OpenAI-compatible provider: point +`base_url` at the provider's endpoint, set `model` to the provider's model +identifier, and put the provider's key in `api_key`. + ## Stop the server Press `Ctrl+C` in the server terminal. diff --git a/docs/configuration.md b/docs/configuration.md index 5cf85a198..485b37fe6 100644 --- a/docs/configuration.md +++ b/docs/configuration.md @@ -124,6 +124,10 @@ Zilliz Cloud endpoint; a Milvus Lite filesystem path is rejected. ### `[llm]` +Any OpenAI-compatible endpoint is supported: set `base_url` to the +provider's endpoint (for example `https://api.deepseek.com` with +`model = "deepseek-flash"`) and `api_key` to that provider's key. + | Field | Type | Default | Required | Description | |---|---|---|---|---| | `model` | string | `"gpt-4.1-mini"` | No | LLM model identifier. |