fix: make llm provider async and add tool toggle

src-tauri/Cargo.lock

@@ -2056,6 +2056,7 @@ checksum = "6373607a59f0be73a39b6fe456b8192fcc3585f602af20751600e974dd455e77"
 name = "living-spec-standalone"
 version = "0.1.0"
 dependencies = [
+ "async-trait",
  "chrono",
  "ignore",
  "reqwest",

src-tauri/Cargo.toml

@@ -28,4 +28,5 @@ walkdir = "2.5.0"
 reqwest = { version = "0.12.28", features = ["json", "blocking"] }
 uuid = { version = "1.19.0", features = ["v4", "serde"] }
 chrono = { version = "0.4.42", features = ["serde"] }
+async-trait = "0.1.89"
 

src-tauri/src/commands/chat.rs

@@ -4,7 +4,7 @@ use crate::llm::types::{
     Message, ModelProvider, Role, ToolCall, ToolDefinition, ToolFunctionDefinition,
 };
 use crate::state::SessionState;
-use serde::{Deserialize, Serialize};
+use serde::Deserialize;
 use serde_json::json;
 use tauri::State;
 
@@ -13,6 +13,7 @@ pub struct ProviderConfig {
     pub provider: String, // "ollama"
     pub model: String,
     pub base_url: Option<String>,
+    pub enable_tools: Option<bool>,
 }
 
 const MAX_TURNS: usize = 10;
@@ -34,7 +35,12 @@ pub async fn chat(
     };
 
     // 2. Define Tools
-    let tools = get_tool_definitions();
+    let tool_defs = get_tool_definitions();
+    let tools = if config.enable_tools.unwrap_or(true) {
+        tool_defs.as_slice()
+    } else {
+        &[]
+    };
 
     // 3. Agent Loop
     let mut current_history = messages.clone();
@@ -49,7 +55,8 @@ pub async fn chat(
 
         // Call LLM
         let response = provider
-            .chat(&config.model, &current_history, &tools)
+            .chat(&config.model, &current_history, tools)
+            .await
             .map_err(|e| format!("LLM Error: {}", e))?;
 
         // Process Response

src-tauri/src/llm/ollama.rs

@@ -1,6 +1,7 @@
 use crate::llm::types::{
     CompletionResponse, FunctionCall, Message, ModelProvider, Role, ToolCall, ToolDefinition,
 };
+use async_trait::async_trait;
 use serde::{Deserialize, Serialize};
 use serde_json::Value;
 
@@ -76,14 +77,15 @@ struct OllamaResponseFunctionCall {
     arguments: Value, // Ollama returns Object, we convert to String for internal storage
 }
 
+#[async_trait]
 impl ModelProvider for OllamaProvider {
-    fn chat(
+    async fn chat(
         &self,
         model: &str,
         messages: &[Message],
         tools: &[ToolDefinition],
     ) -> Result<CompletionResponse, String> {
-        let client = reqwest::blocking::Client::new();
+        let client = reqwest::Client::new();
         let url = format!("{}/api/chat", self.base_url.trim_end_matches('/'));
 
         // Convert domain Messages to Ollama Messages (handling String -> Object args mismatch)
@@ -129,16 +131,18 @@ impl ModelProvider for OllamaProvider {
             .post(&url)
             .json(&request_body)
             .send()
+            .await
             .map_err(|e| format!("Request failed: {}", e))?;
 
         if !res.status().is_success() {
             let status = res.status();
-            let text = res.text().unwrap_or_default();
+            let text = res.text().await.unwrap_or_default();
             return Err(format!("Ollama API error {}: {}", status, text));
         }
 
         let response_body: OllamaResponse = res
             .json()
+            .await
             .map_err(|e| format!("Failed to parse response: {}", e))?;
 
         // Convert Response back to Domain types

src-tauri/src/llm/types.rs

@@ -1,3 +1,4 @@
+use async_trait::async_trait;
 use serde::{Deserialize, Serialize};
 use std::fmt::Debug;
 
@@ -62,8 +63,9 @@ pub struct CompletionResponse {
 }
 
 /// The abstraction for different LLM providers (Ollama, Anthropic, etc.)
+#[async_trait]
 pub trait ModelProvider: Send + Sync {
-    fn chat(
+    async fn chat(
         &self,
         model: &str,
         messages: &[Message],