diff --git a/packages/kilo-docs/lychee.toml b/packages/kilo-docs/lychee.toml
index 0622d7cf074..2f4780b76d6 100644
--- a/packages/kilo-docs/lychee.toml
+++ b/packages/kilo-docs/lychee.toml
@@ -20,6 +20,9 @@ accept = [
 exclude = [
   '\.png$',
   'https://synthetic.new/user-settings/api',
+  # Synthetic's site intermittently returns 503 to automated link checks.
+  '^https?://synthetic\.new/?$',
+  '^https?://synthetic\.new/pricing/?$',
   'https://platform.slack-edge.com/img/add_to_slack.png',
   'https://console.groq.com/pages/models',
   '^https?://([A-Za-z0-9-]+\.)?glama\.ai(/|$)',
diff --git a/packages/kilo-docs/pages/code-with-ai/agents/model-selection.md b/packages/kilo-docs/pages/code-with-ai/agents/model-selection.md
index fae75f4aa3c..d511415d694 100644
--- a/packages/kilo-docs/pages/code-with-ai/agents/model-selection.md
+++ b/packages/kilo-docs/pages/code-with-ai/agents/model-selection.md
@@ -9,6 +9,34 @@ Here's the honest truth about AI model recommendations: by the time I write them
 
 Instead of maintaining a static list that's perpetually behind, we built something better — a real-time leaderboard showing which models Kilo Code users are actually having success with right now.
 
+## Model Routing and Configuration
+
+Kilo's IDE Extension and CLI configurations have four separate, independently configurable model slots:
+
+- **Main model** — the primary model your agent uses for coding tasks, chat, and reasoning. This is what you pick with the model selector, `/models`, or the `model` key in `kilo.jsonc`. See [How to Select and Switch Models](#how-to-select-and-switch-models) for the full precedence order and per-agent config.
+  - The main model is also used for context compaction/summarization and todo-list generation.
+- **Small model** — a lightweight model used for session title generation, commit message generation, and prompt enhancement. Configured with the `small_model` key in `kilo.jsonc`, or the **Small Model** field on the **Settings → Models** tab.
+  - If left unset, Kilo resolves it according to the following logic: 
+    1. Find a small/cheap variant on your current provider (e.g. Haiku on Anthropic, Flash on Gemini).
+    2. Fall back to [`kilo-auto/small`](/docs/gateway/models-and-providers#kilo-autosmall) if the Kilo Gateway provider is authenticated in your session.
+    3. Reuse your main model if you're not authenticated to the Kilo Gateway.
+- **Subagent model** — the default model for subagents launched by the `task` tool. Configured with the `subagent_model` key in `kilo.jsonc`, or the **Subagent Model** field on the **Settings → Models** tab.
+  - If left unset, inherits whichever model the parent agent session is currently using.
+- **Autocomplete model** — the model used for inline code completions as you type. See [Autocomplete: Provider and Model](/docs/code-with-ai/features/autocomplete#provider-and-model) for how to configure it.
+
+### Configuring Local Usage
+
+If you want to use a private/local inference source, you can:
+
+1. Set your **main model** to a non-`kilo-auto` provider — a local model (Ollama, LM Studio) or a direct BYOK provider key.
+2. Explicitly set **`small_model`** to a model on that same provider. If you leave it unset while signed in to Kilo Gateway, it falls back to [`kilo-auto/small`](/docs/gateway/models-and-providers#kilo-autosmall) rather than reusing your main model.
+3. Leave **`subagent_model`** unset — it always inherits your main model rather than defaulting to Kilo Gateway, whether or not you're signed in.
+4. For **autocomplete**, switch the provider to a direct Mistral or Inception BYOK key or disable autocomplete.
+
+{% callout type="note" %}
+You can ensure no inference will go through the Kilo Gateway by logging out of the Kilo Gateway in the IDE extension or TUI.
+{% /callout %}
+
 ## Check the Live Models List
 
 **[👉 See what's working today at kilo.ai/models](https://kilo.ai/models)**
diff --git a/packages/kilo-docs/pages/code-with-ai/gastown/settings.md b/packages/kilo-docs/pages/code-with-ai/gastown/settings.md
index 7512e72e6d6..401ad6eeb32 100644
--- a/packages/kilo-docs/pages/code-with-ai/gastown/settings.md
+++ b/packages/kilo-docs/pages/code-with-ai/gastown/settings.md
@@ -17,15 +17,15 @@ The primary model used by all agents (polecats, refinery, mayor). This affects q
 
 Popular choices:
 - **Kilo Auto Frontier** — highest quality models, best results (recommended)
-- **Kilo Auto Efficient** — cheapest model proven accurate enough for each task, with capability matched to difficulty (minimum for Gas Town)
+- **Kilo Auto Efficient** — cheapest model proven accurate enough for each task, with capability matched to difficulty
 
 ### Role-Specific Models
 
-Override the default model for specific agent roles. By default, all roles use the town-level model. You can override per-role if you want to optimize cost vs quality for different tasks (e.g., a faster model for the mayor, a stronger model for the refinery).
+Override the default model for specific agent roles — **mayor**, **refinery**, and **polecat**. By default, all roles use the town-level default model. You can override per-role if you want to optimize cost vs quality for different tasks (e.g., a faster model for the mayor, a stronger model for the refinery).
 
 ### Small Model
 
-Used for lightweight tasks (classification, routing, summarization). Usually a smaller, cheaper model.
+Used for session title generation and the `explore` subagent inside your town's containers — not general classification, routing, or summarization. This is a lighter-weight model than your default, and it's independent of the `small_model` config used elsewhere in Kilo Code (see [Model Routing and Configuration](/docs/code-with-ai/agents/model-selection#model-routing-and-configuration)). If left unset, it defaults to Claude Haiku.
 
 ## Git & Authentication