clavue 8.32.0 → 9.0.0

package/README.md

@@ -1,4 +1,4 @@
-# Clavue v8.32.0
+# Clavue v9.0.0
 
 ![Clavue cover](https://unpkg.com/clavue/image.png)
 
@@ -65,8 +65,8 @@ npx -y clavue
 Run a specific version with `npx`:
 
 ```bash
-npx -y clavue@8.32.0 --version
-npx -y clavue@8.32.0
+npx -y clavue@9.0.0 --version
+npx -y clavue@9.0.0
 ```
 
 Install globally from npm when you want the `clavue` command to stay available:
@@ -86,7 +86,7 @@ curl -fsSL https://unpkg.com/clavue/install.sh | bash
 Install a specific version globally:
 
 ```bash
-curl -fsSL https://unpkg.com/clavue@8.32.0/install.sh | bash -s -- 8.32.0
+curl -fsSL https://unpkg.com/clavue@9.0.0/install.sh | bash -s -- 9.0.0
 ```
 
 ### Troubleshooting: `ENOTEMPTY` / `bash: clavue: No such file or directory`
@@ -125,17 +125,18 @@ sudo rm -rf "$PREFIX/lib/node_modules/clavue" \
 npm install -g clavue@latest
 ```
 
-## Quick Start: Custom API
+## Quick Start: Provider And Model Setup
 
-Fastest path for custom API users:
+Fastest path for custom API and mixed-model users:
 
 1. Install with `curl -fsSL https://unpkg.com/clavue/install.sh | bash`
 2. Start with `clavue`
-3. At `请选择 API 配置模式`, choose `自定义 API 配置`
-4. At `请选择模型路由模式`, choose the family that matches your models (`OpenAI / GPT`, `Claude`, `Gemini`, `Grok`, or `混合模型`)
-5. At `请选择操作`, choose `1. 添加配置`
-6. Enter a profile name, API base URL, authentication type (`API Key` or `Auth Token`), credential, and primary model. First-run quick add copies the primary model into the other model slots; edit detailed slots later with `clavue provider`
-7. Confirm the saved profile is loaded, then choose `no` at `返回主菜单？` to enter Clavue with the saved default profile
+3. If first-run setup asks for API mode, choose the route you actually use: official login, custom API, CCR proxy, or manual setup later.
+4. Run `/provider` inside Clavue, or `clavue provider` from the shell, to open the provider mode picker.
+5. Choose `组合模式` when different model slots should use different model families or provider routes. Configure every slot explicitly: Primary, Haiku, Sonnet, and Opus.
+6. Choose `单独模式` when one model family and one provider profile should own the whole route.
+7. Choose `完整 API 入口` when you need the full official login / custom API / CCR setup surface.
+8. Run `/model` only for main-session model selection. In combination mode it follows the current `/provider` Primary slot; helper and planning slots remain owned by `/provider`.
 
 Recommended model-slot setup:
 
@@ -230,7 +231,7 @@ Version check:
 ```bash
 clavue --version
 npx -y clavue --version
-npx -y clavue@8.32.0 --version
+npx -y clavue@9.0.0 --version
 # available after a global install and launcher setup
 clavue --version
 ```
@@ -252,7 +253,7 @@ clavue setup-token
 
 ## In-Session Workflows
 
-`/provider` manages saved profiles, current environment state, API URL, credential type, and model-slot routing from one place. Use it when you want to switch, validate, repair, or save the active route.
+`/provider` manages saved profiles, current environment state, API URL, credential type, provider mode, and model-slot routing from one place. With no arguments it starts at the provider mode picker: combination mode for explicit multi-slot routing, single mode for one family/profile, and full API entry for official login, custom API, CCR, or manual setup. Use it when you want to switch, validate, repair, or save the active route.
 
 ```text
 /provider
@@ -268,6 +269,8 @@ clavue setup-token
 /provider clear
 ```
 
+`/model` is the main-session model control. With no arguments it asks whether you want combination mode or single mode. Combination mode keeps the active `/provider` Primary model as the main session model and leaves the other slots under provider control; single mode opens the direct model picker. `/model <model-name>` remains the fast one-command override.
+
 The Mao supervisor ledger is still used internally by delivery gates; user-facing `/mao` and `/codex` command surfaces are hidden while native Team/Agent workflows remain the public agent system.
 
 `/team` inspects real local team state from the active config root instead of giving a generic explanation. By default that is `~/.clavue/teams`, but it follows `CLAVUE_CONFIG_DIR` if you launch clavue against a different config root.