Add Atomic Chat as a first-class local provider.

packages/kilo-docs/lib/nav/ai-providers.ts

@@ -55,6 +55,7 @@ export const AiProvidersNav: NavSection[] = [
     links: [
       { href: "/ai-providers/ollama", children: "Ollama" },
       { href: "/ai-providers/lmstudio", children: "LM Studio" },
+      { href: "/ai-providers/atomic-chat", children: "Atomic Chat" },
       { href: "/ai-providers/vscode-lm", children: "VS Code LM API" },
       {
         href: "/ai-providers/openai-compatible",

packages/kilo-docs/pages/ai-providers/atomic-chat.md

@@ -0,0 +1,112 @@
+---
+title: "Using Atomic Chat with Kilo Code | Local LLMs"
+description: "Run local models in Kilo Code via Atomic Chat's OpenAI-compatible API. Setup for VS Code and the CLI."
+sidebar_label: Atomic Chat
+---
+
+# Using Atomic Chat With Kilo Code
+
+[Kilo Code](https://kilocode.ai/) supports [Atomic Chat](https://atomic.chat/) as a local provider. Atomic Chat runs models on your machine and exposes an OpenAI-compatible API (default `http://127.0.0.1:1337/v1`).
+
+**Website:** [https://atomic.chat/](https://atomic.chat/)  
+**Repository:** [https://github.com/AtomicBot-ai/Atomic-Chat](https://github.com/AtomicBot-ai/Atomic-Chat)
+
+## Prerequisites
+
+1. Install [Atomic Chat](https://atomic.chat/) (macOS or Windows).
+2. Download and load a model in the app.
+3. Enable the **local API server** (default port **1337**).
+4. Confirm the API responds:
+
+```bash
+curl http://127.0.0.1:1337/v1/models
+```
+
+## Configuration in Kilo Code
+
+Kilo Code ships the `@kilocode/plugin-atomic-chat` plugin by default. It **does not** call localhost unless you opt in (see below). When enabled, it discovers models from `GET /v1/models` and can warn if the selected model is not loaded.
+
+**Localhost HTTP runs only when one of these is true:**
+
+- You configure `provider.atomic-chat` in `kilo.jsonc`
+- You set `"model": "atomic-chat/..."` (or per-agent model uses `atomic-chat`)
+- You enable optional auto-detect: `"atomicChat": { "autoDetect": true }` (probes ports **1337** and **1338**)
+
+Otherwise no requests are made to Atomic Chat (suitable for restricted environments).
+
+{% tabs %}
+{% tab label="VSCode" %}
+
+Open **Settings** (gear icon) → **Providers** → **Atomic Chat**. No API key is required for the default local server. Adjust the base URL if Atomic Chat uses a non-default host or port.
+
+{% /tab %}
+{% tab label="CLI" %}
+
+**Config file** (`~/.config/kilo/kilo.jsonc` or `./kilo.jsonc`):
+
+```jsonc
+{
+  "provider": {
+    "atomic-chat": {
+      "options": {
+        "baseURL": "http://127.0.0.1:1337/v1",
+      },
+    },
+  },
+}
+```
+
+Set your default model (use an id from `curl http://127.0.0.1:1337/v1/models`):
+
+```jsonc
+{
+  "model": "atomic-chat/gemma-4-E4B-it-IQ4_XS",
+}
+```
+
+Optional auto-detect without a provider block:
+
+```jsonc
+{
+  "atomicChat": { "autoDetect": true },
+}
+```
+
+To disable the provider entirely, use `disabled_providers: ["atomic-chat"]` or remove `@kilocode/plugin-atomic-chat` from the `plugin` array in your config.
+
+{% /tab %}
+{% /tabs %}
+
+## Custom or unlisted models
+
+If a loaded model does not appear in the picker, register it under `provider.atomic-chat.models`:
+
+```jsonc
+{
+  "model": "atomic-chat/my-local-model",
+  "provider": {
+    "atomic-chat": {
+      "models": {
+        "my-local-model": {
+          "id": "exact-id-from-v1-models",
+          "name": "My Local Model",
+        },
+      },
+    },
+  },
+}
+```
+
+See [Custom Models](/docs/code-with-ai/agents/custom-models) for all model fields.
+
+## Tips
+
+- Prefer capable models with large context windows; agent workflows use long prompts.
+- Keep only the models you need loaded in Atomic Chat to save memory.
+- For embeddings via Atomic Chat, use the **openai-compatible** indexing provider with the same base URL.
+
+## Related
+
+- [LM Studio](/docs/ai-providers/lmstudio)
+- [Ollama](/docs/ai-providers/ollama)
+- [Local models overview](/docs/automate/extending/local-models)

packages/kilo-docs/pages/ai-providers/index.md

@@ -33,6 +33,7 @@ Major AI companies offering powerful models via API:
 
 Run models on your own hardware for privacy and offline use:
 
+- **[Atomic Chat](/docs/ai-providers/atomic-chat)** - Local models with TurboQuant inference and auto-discovery in Kilo Code
 - **[Ollama](/docs/ai-providers/ollama)** - Easy local model management
 - **[LM Studio](/docs/ai-providers/lmstudio)** - Desktop app for local models
 - **[OpenAI Compatible](/docs/ai-providers/openai-compatible)** - Any OpenAI-compatible endpoint

packages/kilo-docs/pages/automate/extending/local-models.md

@@ -5,7 +5,7 @@ description: "Run AI models locally with Kilo Code"
 
 # Using Local Models
 
-Kilo Code supports running language models locally on your own machine using [Ollama](https://ollama.com/) and [LM Studio](https://lmstudio.ai/). This offers several advantages:
+Kilo Code supports running language models locally on your own machine using [Ollama](https://ollama.com/), [LM Studio](https://lmstudio.ai/), and [Atomic Chat](https://atomic.chat/). This offers several advantages:
 
 - **Privacy:** Your code and data never leave your computer.
 - **Offline Access:** You can use Kilo Code even without an internet connection.
@@ -21,23 +21,23 @@ Kilo Code supports running language models locally on your own machine using [Ol
 
 ## Supported Local Model Providers
 
-Kilo Code currently supports two main local model providers:
+Kilo Code supports several local model providers:
 
 1.  **Ollama:** A popular open-source tool for running large language models locally. It supports a wide range of models.
-2.  **LM Studio:** A user-friendly desktop application that simplifies the process of downloading, configuring, and running local models. It also provides a local server that emulates the OpenAI API.
+2.  **LM Studio:** A user-friendly desktop application that simplifies downloading and running local models, with a local server that emulates the OpenAI API.
+3.  **[Atomic Chat](https://atomic.chat/):** Open-source local AI with TurboQuant-optimized inference, a built-in chat UI, and an OpenAI-compatible API on port **1337**. Kilo Code can discover loaded models when you opt in (`provider.atomic-chat`, `atomicChat.autoDetect`, or an `atomic-chat/...` model).
 
 ## Setting Up Local Models
 
 For detailed setup instructions, see:
 
 - [Setting up Ollama](/docs/ai-providers/ollama)
 - [Setting up LM Studio](/docs/ai-providers/lmstudio)
-
-Both providers offer similar capabilities but with different user interfaces and workflows. Ollama provides more control through its command-line interface, while LM Studio offers a more user-friendly graphical interface.
+- [Setting up Atomic Chat](/docs/ai-providers/atomic-chat)
 
 ## Troubleshooting
 
-- **"No connection could be made because the target machine actively refused it":** This usually means that the Ollama or LM Studio server isn't running, or is running on a different port/address than Kilo Code is configured to use. Double-check the Base URL setting.
+- **"No connection could be made because the target machine actively refused it":** This usually means that Atomic Chat, Ollama, or LM Studio isn't running, or is on a different port than Kilo Code expects (Atomic Chat: `http://127.0.0.1:1337/v1`, LM Studio: `http://127.0.0.1:1234/v1`, Ollama: `http://127.0.0.1:11434`). Double-check the Base URL setting.
 
 - **Slow Response Times:** Local models can be slower than cloud-based models, especially on less powerful hardware. If performance is an issue, try using a smaller model.
 

packages/kilo-vscode/webview-ui/src/components/settings/ProviderConnectDialog.tsx

@@ -12,6 +12,11 @@ import { useLanguage } from "../../context/language"
 import { useProvider } from "../../context/provider"
 import { useVSCode } from "../../context/vscode"
 import { createProviderAction } from "../../utils/provider-action"
+import {
+  ATOMIC_CHAT_PROVIDER_KEY,
+  isLocalProviderOptionalApiKey,
+  LOCAL_PROVIDER_API_KEY_PLACEHOLDER,
+} from "../../utils/local-providers"
 
 interface ProviderConnectDialogProps {
   providerID: string
@@ -265,7 +270,7 @@ const ProviderConnectDialog: Component<ProviderConnectDialogProps> = (props) =>
           <For each={methods()}>
             {(item, index) => (
               <Button variant="secondary" size="large" onClick={() => selectMethod(index())}>
-                {item.type === "api" ? language.t("provider.connect.method.apiKey") : item.label}
+                {item.type === "api" ? item.label || language.t("provider.connect.method.apiKey") : item.label}
               </Button>
             )}
           </For>
@@ -283,10 +288,29 @@ const ProviderConnectDialog: Component<ProviderConnectDialogProps> = (props) =>
     const [value, setValue] = createSignal("")
     const [fields, setFields] = createStore<Record<string, string>>({})
     const prompts = createMemo(() => method()?.prompts?.filter((prompt) => visible(prompt, fields)) ?? [])
+    const apiKeyOptional = () => isLocalProviderOptionalApiKey(props.providerID)
+
+    function apiKeyDescription() {
+      if (props.providerID === ATOMIC_CHAT_PROVIDER_KEY) {
+        return language.t("provider.connect.atomicChat.description")
+      }
+      if (apiKeyOptional()) {
+        return language.t("provider.connect.apiKey.description.local", { provider: name() })
+      }
+      return language.t("provider.connect.apiKey.description", { provider: name() })
+    }
+
+    function apiKeyLabel() {
+      if (apiKeyOptional()) {
+        return language.t("provider.connect.apiKey.label.optional", { provider: name() })
+      }
+      return language.t("provider.connect.apiKey.label", { provider: name() })
+    }
 
     function submit(e: SubmitEvent) {
       e.preventDefault()
-      const apiKey = value().trim()
+      const trimmed = value().trim()
+      const apiKey = trimmed || (apiKeyOptional() ? LOCAL_PROVIDER_API_KEY_PLACEHOLDER : "")
       if (!apiKey) {
         setState({ ...state, error: language.t("provider.connect.apiKey.required"), field: "apiKey" })
         return
@@ -313,14 +337,16 @@ const ProviderConnectDialog: Component<ProviderConnectDialogProps> = (props) =>
         style={{ display: "flex", "flex-direction": "column", gap: "16px" }}
         onSubmit={submit}
       >
-        <div class="provider-connect-body">
-          {language.t("provider.connect.apiKey.description", { provider: name() })}
-        </div>
+        <div class="provider-connect-body">{apiKeyDescription()}</div>
         <TextField
           autofocus
           type="password"
-          label={language.t("provider.connect.apiKey.label", { provider: name() })}
-          placeholder={language.t("provider.connect.apiKey.placeholder")}
+          label={apiKeyLabel()}
+          placeholder={
+            apiKeyOptional()
+              ? language.t("provider.connect.apiKey.placeholder.optional")
+              : language.t("provider.connect.apiKey.placeholder")
+          }
           value={value()}
           onChange={setValue}
           validationState={state.field === "apiKey" ? "invalid" : undefined}

packages/kilo-vscode/webview-ui/src/i18n/en.ts

@@ -136,8 +136,14 @@ export const dict = {
   "provider.connect.status.failed": "Authorization failed: {{error}}",
   "provider.connect.apiKey.description":
     "Enter your {{provider}} API key to connect your account and use {{provider}} models in Kilo.",
+  "provider.connect.apiKey.description.local":
+    "Connect to your local {{provider}} server. Leave the API key empty if the server does not require one (default for localhost).",
+  "provider.connect.atomicChat.description":
+    "Connect to Atomic Chat on your machine (default http://127.0.0.1:1337). No API key is required for the local server — start Atomic Chat, load a model, then connect.",
   "provider.connect.apiKey.label": "{{provider}} API key",
+  "provider.connect.apiKey.label.optional": "{{provider}} API key (optional)",
   "provider.connect.apiKey.placeholder": "API key",
+  "provider.connect.apiKey.placeholder.optional": "Leave empty for local server",
   "provider.connect.apiKey.required": "API key is required",
   "provider.connect.prompt.required": "{{field}} is required",
   "provider.connect.azure.endpointType.label": "Select Azure endpoint configuration",