From 0f0f250bc7a15155773a6d11898aaf1830155573 Mon Sep 17 00:00:00 2001 From: DaoXE Team Date: Wed, 15 Jul 2026 00:48:14 +0800 Subject: [PATCH] docs(tutorials): add DaoXE OpenAI-compatible connection guide Community tutorial for connecting Open WebUI to DaoXE via Admin OpenAI connections (https://daoxe.com/v1), with optional Direct Connections notes and account-scoped model discovery. --- docs/tutorials/integrations/index.md | 1 + .../integrations/llm-providers/daoxe.md | 67 +++++++++++++++++++ 2 files changed, 68 insertions(+) create mode 100644 docs/tutorials/integrations/llm-providers/daoxe.md diff --git a/docs/tutorials/integrations/index.md b/docs/tutorials/integrations/index.md index 8486d6794..15e3bcc13 100644 --- a/docs/tutorials/integrations/index.md +++ b/docs/tutorials/integrations/index.md @@ -16,6 +16,7 @@ These community-contributed guides cover specific integration scenarios with ste | Tutorial | What you'll achieve | Details | |----------|-------------------|---------| | [Azure OpenAI (Entra ID)](./llm-providers/azure-openai) | Keyless authentication to Azure OpenAI via Entra ID | 👤 Admin · ⏱️ 30 to 60 min | +| [DaoXE](./llm-providers/daoxe) | Connect Open WebUI to DaoXE as an OpenAI-compatible gateway | 👤 Admin · ⏱️ 5 min | | [DeepSeek R1 Dynamic](./llm-providers/deepseekr1-dynamic) | Run the full 671B DeepSeek-R1 model via llama.cpp | 👤 Developer · ⏱️ 45 min | | [Intel GPU (IPEX-LLM)](./llm-providers/ipex_llm) | Accelerate Ollama with IPEX-LLM on Intel GPUs | 👤 Developer · ⏱️ 20 min | diff --git a/docs/tutorials/integrations/llm-providers/daoxe.md b/docs/tutorials/integrations/llm-providers/daoxe.md new file mode 100644 index 000000000..6ca6791f6 --- /dev/null +++ b/docs/tutorials/integrations/llm-providers/daoxe.md @@ -0,0 +1,67 @@ +--- +sidebar_position: 40 +title: "DaoXE" +--- + +:::warning + +This tutorial is a community contribution and is not supported by the Open WebUI team. It serves only as a demonstration on how to customize Open WebUI for your specific use case. Want to contribute? Check out the [contributing tutorial](/contributing). + +::: + +# Connect Open WebUI to DaoXE + +**DaoXE** is a multi-model multi-protocol AI API gateway. For Open WebUI, use the **OpenAI-compatible** Chat Completions path: set the connection URL to `https://daoxe.com/v1`, add your API key, and select a model ID from your account. + +:::note +DaoXE is **not available in mainland China**. Model availability is account-scoped (there is no fixed public model list). Setup examples: [DaoXE-AI](https://github.com/seven7763/DaoXE-AI). +::: + +## Prerequisites + +- An Open WebUI deployment with admin access to **Admin Settings → Connections** +- A DaoXE account and API key from [daoxe.com](https://daoxe.com) +- At least one model enabled for that key (check the DaoXE dashboard or `GET /v1/models`) + +## Admin connection (recommended) + +1. Open Open WebUI and go to **Admin Settings → Connections → OpenAI**. +2. Click **Add Connection** (or the **+** icon under OpenAI API connections). +3. Fill in: + +| Setting | Value | +|---|---| +| **URL** | `https://daoxe.com/v1` | +| **API Key** | Your DaoXE API key | +| **Model IDs (Filter)** | Optional. Leave empty to use authenticated `/models` discovery, or add specific model IDs from your account | + +4. Save the connection. If verification succeeds, models for that key appear in the model selector. +5. Start a chat and pick a DaoXE model from the dropdown. + +:::tip +Open WebUI verifies connections with `GET /models` using a Bearer token. DaoXE supports authenticated model listing for the key you provide. If your catalog is large, add a short allowlist under **Model IDs (Filter)** so the selector stays manageable. +::: + +:::warning No trailing slash +Use exactly `https://daoxe.com/v1` (no trailing slash). A trailing slash can break the `/models` path Open WebUI appends. +::: + +## Direct Connections (optional) + +If the admin has enabled [Direct Connections](/features/chat-conversations/direct-connections) (`ENABLE_DIRECT_CONNECTIONS=true`): + +1. Go to **User Settings → Connections**. +2. Add a connection with the same **Base URL** (`https://daoxe.com/v1`) and your personal API key. +3. Save. Requests then go from the browser to DaoXE (CORS must allow your Open WebUI origin). + +## Notes + +- This path is OpenAI-compatible **Chat Completions** (`/v1/chat/completions`). DaoXE also exposes other protocols (for example OpenAI Responses and Anthropic Messages) outside this Open WebUI connection type. +- Use model IDs from **your** DaoXE account only; do not hard-code a static third-party price/model list. +- Prefer least-privilege keys for multi-user deployments. + +## Learn more + +- DaoXE: [https://daoxe.com](https://daoxe.com) +- Examples: [https://github.com/seven7763/DaoXE-AI](https://github.com/seven7763/DaoXE-AI) +- Open WebUI OpenAI-compatible guide: [OpenAI-Compatible Providers](/getting-started/quick-start/connect-a-provider/starting-with-openai-compatible)