From 4420f194161364638c3b91fa251218eab802f58c Mon Sep 17 00:00:00 2001 From: yanalialiuk Date: Tue, 30 Jun 2026 02:55:10 +0300 Subject: [PATCH 1/2] docs: add Atomic Chat as local OpenAI-compatible provider Co-authored-by: Cursor --- admins/ai_models/atomic_chat.mdx | 61 ++++++++++++++++++++++++++++++++ admins/ai_models/overview.mdx | 6 ++-- docs.json | 1 + 3 files changed, 66 insertions(+), 2 deletions(-) create mode 100644 admins/ai_models/atomic_chat.mdx diff --git a/admins/ai_models/atomic_chat.mdx b/admins/ai_models/atomic_chat.mdx new file mode 100644 index 00000000..dbd15240 --- /dev/null +++ b/admins/ai_models/atomic_chat.mdx @@ -0,0 +1,61 @@ +--- +title: "Atomic Chat" +description: "Using Atomic Chat with Onyx" +icon: "message" +--- + +import NavigateToAiModels from "/snippets/navigate-to-ai-models.mdx"; + +## Guide + +Configure Onyx to use models served locally by [Atomic Chat](https://github.com/AtomicBot-ai/Atomic-Chat). + +Atomic Chat exposes an **OpenAI-compatible API** at `http://localhost:1337/v1`. Load a model in the Atomic Chat app, then connect Onyx through a custom OpenAI-compatible provider. + + + + Download Atomic Chat from the [releases page](https://github.com/AtomicBot-ai/Atomic-Chat/releases) or [atomic.chat](https://atomic.chat/). + + Open the app and load the model you want to use. The local API server starts automatically on port `1337`. + + Verify the server is running: + + ```bash + curl http://127.0.0.1:1337/v1/models + ``` + + + **Running Onyx in Docker on the same machine as Atomic Chat?** + + Use `http://host.docker.internal:1337/v1` as the **Base URL** instead of `http://localhost:1337/v1`. + Inside a container, `localhost` refers to the container itself, not your host. + + + + + + + Select **Add Custom LLM Provider** from the available providers. + + Give your provider a **Display Name** (for example, `Atomic Chat`). + + Set **Provider Name** to `openai`. Atomic Chat implements the OpenAI API surface that LiteLLM expects for chat completions. + + Set **Base URL** to `http://127.0.0.1:1337/v1` (or `http://host.docker.internal:1337/v1` if Onyx runs in Docker). + + If Atomic Chat requires an API key in your setup, enter any non-empty placeholder such as `atomic`. + + + + In **Model Configurations**, add the model name exactly as shown in Atomic Chat (for example, the loaded GGUF or Hugging Face model ID). + + Click **Fetch Available Models** if your Onyx version supports discovery against OpenAI-compatible `/models` endpoints. + + Choose which models are visible to your users and whether the provider is public or private. + + + +## Related resources + +- [Atomic Chat API docs](https://github.com/AtomicBot-ai/Atomic-Chat#-use-it-as-an-api) +- [Custom Inference Provider](/admins/ai_models/custom_inference_provider) for other OpenAI-compatible gateways diff --git a/admins/ai_models/overview.mdx b/admins/ai_models/overview.mdx index ff0c9343..d0abbdb5 100644 --- a/admins/ai_models/overview.mdx +++ b/admins/ai_models/overview.mdx @@ -41,8 +41,8 @@ which models are visible, and which models should be the default or fast option These options let you run open-weight models on your own infrastructure or local hardware. - Onyx includes built-in integrations for [Ollama](/admins/ai_models/ollama) - and [LM Studio](/admins/ai_models/lm_studio). This is a good fit when your team needs local development workflows, + Onyx includes built-in integrations for [Ollama](/admins/ai_models/ollama), + [LM Studio](/admins/ai_models/lm_studio), and [Atomic Chat](/admins/ai_models/atomic_chat). This is a good fit when your team needs local development workflows, tighter data residency, or lower per-token costs. @@ -99,6 +99,8 @@ which models are visible, and which models should be the default or fast option + + diff --git a/docs.json b/docs.json index 78c53962..c7472fa2 100644 --- a/docs.json +++ b/docs.json @@ -152,6 +152,7 @@ "admins/ai_models/anthropic", "admins/ai_models/ollama", "admins/ai_models/lm_studio", + "admins/ai_models/atomic_chat", "admins/ai_models/azure_openai", "admins/ai_models/bedrock", "admins/ai_models/google_ai", From 179341b0bf69e7c41ecd3c94f5efb7d531079d5c Mon Sep 17 00:00:00 2001 From: yanalialiuk Date: Tue, 30 Jun 2026 02:55:24 +0300 Subject: [PATCH 2/2] docs: clarify Atomic Chat uses custom provider setup Co-authored-by: Cursor --- admins/ai_models/overview.mdx | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/admins/ai_models/overview.mdx b/admins/ai_models/overview.mdx index d0abbdb5..60c482cb 100644 --- a/admins/ai_models/overview.mdx +++ b/admins/ai_models/overview.mdx @@ -41,8 +41,8 @@ which models are visible, and which models should be the default or fast option These options let you run open-weight models on your own infrastructure or local hardware. - Onyx includes built-in integrations for [Ollama](/admins/ai_models/ollama), - [LM Studio](/admins/ai_models/lm_studio), and [Atomic Chat](/admins/ai_models/atomic_chat). This is a good fit when your team needs local development workflows, + Onyx includes built-in integrations for [Ollama](/admins/ai_models/ollama) + and [LM Studio](/admins/ai_models/lm_studio). You can also connect [Atomic Chat](/admins/ai_models/atomic_chat) and other local runtimes through an OpenAI-compatible endpoint. This is a good fit when your team needs local development workflows, tighter data residency, or lower per-token costs.