diff --git a/Cargo.toml b/Cargo.toml
index c45d272..80fb80d 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -2,7 +2,7 @@
 name = "iAs"
 version = "0.1.0"
 edition = "2024"
-description = "微信 AI 智能助手 — 通过 DeepSeek LLM 实现 AI 自动回复"
+description = "AI 智能助手"
 
 [dependencies]
 # ── 异步运行时 ──
@@ -51,4 +51,3 @@ scraper = "0.27.0"                                             # HTML 解析（f
 [[bin]]
 name = "ias"
 path = "src/main.rs"
-
diff --git a/crates/ias-fetch/Cargo.toml b/crates/ias-fetch/Cargo.toml
deleted file mode 100644
index f2e3df8..0000000
--- a/crates/ias-fetch/Cargo.toml
+++ /dev/null
@@ -1,10 +0,0 @@
-[package]
-name = "ias-fetch"
-version = "0.1.0"
-edition = "2024"
-
-[dependencies]
-chromiumoxide = { version = "0.7", default-features = false, features = ["tokio-runtime", "bytes"] }
-futures = "0.3"
-serde_json = "1.0"
-tokio = { version = "1.0", features = ["full"] }
diff --git a/crates/ias-fetch/src/main.rs b/crates/ias-fetch/src/main.rs
deleted file mode 100644
index f500959..0000000
--- a/crates/ias-fetch/src/main.rs
+++ /dev/null
@@ -1,143 +0,0 @@
-//! ias-fetch — 通过 headless Chrome 提取网页正文
-//!
-//! 启动无头 Chrome 渲染页面 JS，取渲染后的纯文本输出到 stdout。
-//! 专供 JSON 工具 `${cmd:ias-fetch {{url}}}` 调用。
-//!
-//! 用法:
-//!   ias-fetch <url>                    # 默认：取 body.innerText
-//!   ias-fetch --raw <url>              # 返回完整 HTML
-//!   ias-fetch --timeout 15 <url>       # 自定义超时（默认 15s）
-
-use futures::StreamExt;
-use std::time::Duration;
-
-use chromiumoxide::{Browser, BrowserConfig, Page};
-
-const DEFAULT_TIMEOUT: u64 = 15;
-const MAX_OUTPUT_CHARS: usize = 10_000;
-
-#[tokio::main]
-async fn main() {
-    let args: Vec<String> = std::env::args().collect();
-
-    let mut raw_mode = false;
-    let mut timeout_secs = DEFAULT_TIMEOUT;
-    let mut url: Option<String> = None;
-
-    let mut i = 1;
-    while i < args.len() {
-        match args[i].as_str() {
-            "--raw" => raw_mode = true,
-            "--timeout" => {
-                i += 1;
-                timeout_secs = args.get(i).and_then(|s| s.parse().ok()).unwrap_or(DEFAULT_TIMEOUT);
-            }
-            s if s.starts_with("http://") || s.starts_with("https://") => {
-                url = Some(s.to_string());
-            }
-            s if !s.starts_with('-') => url = Some(s.to_string()),
-            _ => {}
-        }
-        i += 1;
-    }
-
-    let url = match url {
-        Some(u) => {
-            if u.starts_with("http://") || u.starts_with("https://") {
-                u
-            } else {
-                format!("https://{}", u)
-            }
-        }
-        None => {
-            eprintln!("用法: ias-fetch [--raw] [--timeout N] <url>");
-            std::process::exit(1);
-        }
-    };
-
-    let timeout = Duration::from_secs(timeout_secs);
-
-    match tokio::time::timeout(timeout, fetch_page(&url, raw_mode)).await {
-        Ok(Ok(output)) => {
-            if output.is_empty() {
-                eprintln!("页面无内容");
-                std::process::exit(1);
-            }
-            println!("{}", output);
-        }
-        Ok(Err(e)) => {
-            eprintln!("{}", e);
-            std::process::exit(1);
-        }
-        Err(_) => {
-            eprintln!("超时 ({}s): {}", timeout_secs, url);
-            std::process::exit(1);
-        }
-    }
-}
-
-async fn fetch_page(url: &str, raw_mode: bool) -> Result<String, String> {
-    let (mut browser, mut handler) = Browser::launch(
-        BrowserConfig::builder()
-            .build()
-            .map_err(|e| format!("Chrome 配置失败: {}", e))?,
-    )
-    .await
-    .map_err(|e| format!("启动 Chrome 失败: {}\n请确保已安装 chromium 或 google-chrome", e))?;
-
-    let handle = tokio::spawn(async move {
-        while let Some(_event) = handler.next().await {
-            // 持续消费 handler 事件
-        }
-    });
-
-    // new_page 接受 URL，自动导航并等待加载
-    let page = browser
-        .new_page(url)
-        .await
-        .map_err(|e| format!("导航到 {} 失败: {}", url, e))?;
-
-    // 额外等待 JS 渲染
-    tokio::time::sleep(Duration::from_secs(2)).await;
-
-    // 提取内容
-    let content = if raw_mode {
-        page.content().await.map_err(|e| format!("获取 HTML 失败: {}", e))?
-    } else {
-        get_inner_text(&page).await?
-    };
-
-    // 清理
-    browser.close().await.ok();
-    handle.await.ok();
-
-    // 截断
-    if content.chars().count() > MAX_OUTPUT_CHARS {
-        Ok(format!(
-            "{}…（已截断，完整内容 {} 字符）",
-            content.chars().take(MAX_OUTPUT_CHARS).collect::<String>(),
-            content.chars().count()
-        ))
-    } else {
-        Ok(content)
-    }
-}
-
-/// 通过 evaluate JS 获取 body.innerText
-async fn get_inner_text(page: &Page) -> Result<String, String> {
-    let result = page
-        .evaluate("document.body?.innerText || document.documentElement?.innerText || ''")
-        .await
-        .map_err(|e| format!("提取正文失败: {}", e))?;
-
-    let value: serde_json::Value = result
-        .into_value()
-        .map_err(|e| format!("解析 evaluate 结果失败: {}", e))?;
-    match value {
-        serde_json::Value::String(s) => {
-            let cleaned: Vec<&str> = s.lines().map(|l| l.trim()).filter(|l| !l.is_empty()).collect();
-            Ok(cleaned.join("\n"))
-        }
-        _ => Ok(String::new()),
-    }
-}