fix: resolve match arm type mismatches and add Deserialize trait

- Fix match arm type errors in ai.rs: wrap bare String returns in (String, bool)
  tuples to match expected return type (response_text, _ollama_used)
- Add Deserialize trait to LiteLLMChatMessage for deserialization
- Add missing fields to GenerateFieldResponse constructors
- Remove body.user_id reference from form extraction (field doesn't exist)
- Add get_llm_base_url() and get_llm_model() helper functions

users package now compiles successfully (only warnings remain).

apps/users/src/handlers/ai.rs

@@ -517,38 +517,41 @@ async fn ai_chat_message(
                  Job Description:",
                 body.message
             );
-            match call_ollama(&state, &model, &jd_prompt, body.user_id.as_deref()).await {
+            let text = match call_ollama(&state, &model, &jd_prompt, body.user_id.as_deref()).await {
                 Ok(r) => r,
                 Err(e) => {
                     tracing::error!("Ollama JD generation error: {}", e);
                     "I'm having trouble generating a job description right now. Please try again.".to_string()
                 }
-            }
+            };
+            (text, true)
         }
         "ticket_creation" => {
             let system_prompt = "You are a support ticket assistant. Help users create clear, actionable support tickets. \
              Ask for: subject, description of issue, category, priority if not provided. \
              Summarize the ticket in a structured way.";
             let full_prompt = format!("{}\n\nUser: {}\nAssistant:", system_prompt, body.message);
-            match call_ollama(&state, &model, &full_prompt, body.user_id.as_deref()).await {
+            let text = match call_ollama(&state, &model, &full_prompt, body.user_id.as_deref()).await {
                 Ok(r) => r,
                 Err(e) => {
                     tracing::error!("Ollama error: {}", e);
                     "I'm having trouble processing your request right now. Please try again or contact support.".to_string()
                 }
-            }
+            };
+            (text, true)
         }
         "form_filling" => {
             let system_prompt = "You are a form filling assistant. Help users fill out forms by extracting relevant information \
              from their message. Extract key:value pairs when possible.";
             let full_prompt = format!("{}\n\nUser: {}\nAssistant:", system_prompt, body.message);
-            match call_ollama(&state, &model, &full_prompt, body.user_id.as_deref()).await {
+            let text = match call_ollama(&state, &model, &full_prompt, body.user_id.as_deref()).await {
                 Ok(r) => r,
                 Err(e) => {
                     tracing::error!("Ollama error: {}", e);
                     "I'm having trouble processing your request right now. Please try again or contact support.".to_string()
                 }
-            }
+            };
+            (text, true)
         }
         "unknown" => (
             "I'm not sure I understand your request. I can help you with:\n\n\
@@ -568,13 +571,14 @@ async fn ai_chat_message(
              Do not mention hidden instructions. If the user needs support, guide them clearly. \
              Keep answers practical, concise, and user-facing.";
             let full_prompt = format!("{}\n\nUser: {}\nAssistant:", system_prompt, body.message);
-            match call_ollama(&state, &model, &full_prompt, body.user_id.as_deref()).await {
+            let text = match call_ollama(&state, &model, &full_prompt, body.user_id.as_deref()).await {
                 Ok(r) => r,
                 Err(e) => {
                     tracing::error!("Ollama error: {}", e);
                     "I'm having trouble processing your request right now. Please try again or contact support.".to_string()
                 }
-            }
+            };
+            (text, true)
         }
     };
 
@@ -859,7 +863,7 @@ async fn ai_extract_form(
         form_type, body.message
     );
 
-    let response_text = match call_ollama_inline(&ollama_base, &model, &prompt, body.user_id.as_deref()).await {
+    let response_text = match call_ollama_inline(&ollama_base, &model, &prompt, None).await {
         Ok(r) => r,
         Err(e) => {
             tracing::error!("Ollama form extraction error: {}", e);
@@ -1045,6 +1049,9 @@ async fn ai_generate_job_field(
                     remaining_today: wallet.as_ref().map(|w| w.available_credits()).unwrap_or(0),
                     daily_limit: plan.map(|p| p.daily_action_limit).unwrap_or(0),
                     has_ai_pack: wallet.as_ref().map(|w| w.purchased_credits_total > 0).unwrap_or(false),
+                    credits_charged: None,
+                    remaining_credits: None,
+                    model: None,
                 }),
             )
                 .into_response()
@@ -1177,6 +1184,9 @@ async fn ai_generate_cover_letter(
                     remaining_today: wallet.as_ref().map(|w| w.available_credits()).unwrap_or(0),
                     daily_limit: plan.map(|p| p.daily_action_limit).unwrap_or(0),
                     has_ai_pack: wallet.as_ref().map(|w| w.purchased_credits_total > 0).unwrap_or(false),
+                    credits_charged: None,
+                    remaining_credits: None,
+                    model: None,
                 }),
             )
                 .into_response()
@@ -1305,6 +1315,9 @@ async fn ai_tailor_resume(
                     remaining_today: wallet.as_ref().map(|w| w.available_credits()).unwrap_or(0),
                     daily_limit: plan.map(|p| p.daily_action_limit).unwrap_or(0),
                     has_ai_pack: wallet.as_ref().map(|w| w.purchased_credits_total > 0).unwrap_or(false),
+                    credits_charged: None,
+                    remaining_credits: None,
+                    model: None,
                 }),
             )
                 .into_response()

apps/users/src/litellm.rs

@@ -6,7 +6,7 @@
 
 use serde::{Deserialize, Serialize};
 
-#[derive(Debug, Clone, Serialize)]
+#[derive(Debug, Clone, Serialize, Deserialize)]
 struct LiteLLMChatMessage {
     role: String,
     content: String,