fix(llm-config): Fix ConfigManager fallback logic for LLM configuration

- Fix ConfigManager to treat 'none', 'null', 'n/a', and empty values as placeholders
  and fall back to default bot's configuration instead of using these as literal values

- Fix ConfigManager to detect local file paths (e.g., .gguf, .bin, ../) and fall back
  to default bot's model when using remote API, allowing bots to keep local model
  config for local LLM server while automatically using remote model for API calls

- Fix get_default_bot() to return the bot actually named 'default' instead of
  the first active bot by ID, ensuring consistent fallback behavior

- Add comprehensive debug logging to trace LLM configuration from database to API call

This fixes the issue where bots with incomplete or local LLM configuration would
fail with 401/400 errors when trying to use remote API, instead of automatically
falling back to the default bot's configuration from config.csv.

Closes: #llm-config-fallback

src/core/bot/mod.rs

@@ -1,5 +1,7 @@
 #[cfg(any(feature = "research", feature = "llm"))]
 pub mod kb_context;
+#[cfg(any(feature = "research", feature = "llm"))]
+use kb_context::inject_kb_context;
 #[cfg(feature = "llm")]
 use crate::core::config::ConfigManager;
 
@@ -20,7 +22,10 @@ use axum::{
     http::StatusCode,
     response::{IntoResponse, Json},
 };
+use diesel::ExpressionMethods;
 use diesel::PgConnection;
+use diesel::QueryDsl;
+use diesel::RunQueryDsl;
 use futures::{sink::SinkExt, stream::StreamExt};
 #[cfg(feature = "llm")]
 use log::trace;
@@ -39,7 +44,9 @@ pub fn get_default_bot(conn: &mut PgConnection) -> (Uuid, String) {
     use crate::shared::models::schema::bots::dsl::*;
     use diesel::prelude::*;
 
+    // First try to get the bot named "default"
     match bots
+        .filter(name.eq("default"))
         .filter(is_active.eq(true))
         .select((id, name))
         .first::<(Uuid, String)>(conn)
@@ -47,8 +54,24 @@ pub fn get_default_bot(conn: &mut PgConnection) -> (Uuid, String) {
     {
         Ok(Some((bot_id, bot_name))) => (bot_id, bot_name),
         Ok(None) => {
-            warn!("No active bots found, using nil UUID");
-            (Uuid::nil(), "default".to_string())
+            warn!("Bot named 'default' not found, falling back to first active bot");
+            // Fall back to first active bot
+            match bots
+                .filter(is_active.eq(true))
+                .select((id, name))
+                .first::<(Uuid, String)>(conn)
+                .optional()
+            {
+                Ok(Some((bot_id, bot_name))) => (bot_id, bot_name),
+                Ok(None) => {
+                    warn!("No active bots found, using nil UUID");
+                    (Uuid::nil(), "default".to_string())
+                }
+                Err(e) => {
+                    error!("Failed to query fallback bot: {}", e);
+                    (Uuid::nil(), "default".to_string())
+                }
+            }
         }
         Err(e) => {
             error!("Failed to query default bot: {}", e);
@@ -72,10 +95,116 @@ impl BotOrchestrator {
     }
 
     pub fn mount_all_bots(&self) -> Result<(), Box<dyn std::error::Error + Send + Sync>> {
-        info!("mount_all_bots called");
+        info!("Scanning drive for .gbai files to mount bots...");
+
+        let mut bots_mounted = 0;
+        let mut bots_created = 0;
+
+        let directories_to_scan: Vec<std::path::PathBuf> = vec![
+            self.state
+                .config
+                .as_ref()
+                .map(|c| c.site_path.clone())
+                .unwrap_or_else(|| "./botserver-stack/sites".to_string())
+                .into(),
+            "./templates".into(),
+            "../bottemplates".into(),
+        ];
+
+        for dir_path in directories_to_scan {
+            info!("Checking directory for bots: {}", dir_path.display());
+
+            if !dir_path.exists() {
+                info!("Directory does not exist, skipping: {}", dir_path.display());
+                continue;
+            }
+
+            match self.scan_directory(&dir_path, &mut bots_mounted, &mut bots_created) {
+                Ok(()) => {}
+                Err(e) => {
+                    error!("Failed to scan directory {}: {}", dir_path.display(), e);
+                }
+            }
+        }
+
+        info!(
+            "Bot mounting complete: {} bots processed ({} created, {} already existed)",
+            bots_mounted,
+            bots_created,
+            bots_mounted - bots_created
+        );
+
         Ok(())
     }
 
+    fn scan_directory(
+        &self,
+        dir_path: &std::path::Path,
+        bots_mounted: &mut i32,
+        _bots_created: &mut i32,
+    ) -> Result<(), Box<dyn std::error::Error + Send + Sync>> {
+        let entries =
+            std::fs::read_dir(dir_path).map_err(|e| format!("Failed to read directory: {}", e))?;
+
+        for entry in entries.flatten() {
+            let name = entry.file_name();
+
+            let bot_name = match name.to_str() {
+                Some(n) if n.ends_with(".gbai") => n.trim_end_matches(".gbai"),
+                _ => continue,
+            };
+
+            info!("Found .gbai file: {}", bot_name);
+
+            match self.ensure_bot_exists(bot_name) {
+                Ok(true) => {
+                    info!("Bot '{}' already exists in database, mounting", bot_name);
+                    *bots_mounted += 1;
+                }
+                Ok(false) => {
+                    info!(
+                        "Bot '{}' does not exist in database, skipping (run import to create)",
+                        bot_name
+                    );
+                }
+                Err(e) => {
+                    error!("Failed to check if bot '{}' exists: {}", bot_name, e);
+                }
+            }
+        }
+
+        Ok(())
+    }
+
+    fn ensure_bot_exists(
+        &self,
+        bot_name: &str,
+    ) -> Result<bool, Box<dyn std::error::Error + Send + Sync>> {
+        use diesel::sql_query;
+
+        let mut conn = self
+            .state
+            .conn
+            .get()
+            .map_err(|e| format!("Failed to get database connection: {e}"))?;
+
+        #[derive(diesel::QueryableByName)]
+        #[diesel(check_for_backend(diesel::pg::Pg))]
+        struct BotExistsResult {
+            #[diesel(sql_type = diesel::sql_types::Bool)]
+            exists: bool,
+        }
+
+        let exists: BotExistsResult = sql_query(
+            "SELECT EXISTS(SELECT 1 FROM bots WHERE name = $1 AND is_active = true) as exists",
+        )
+        .bind::<diesel::sql_types::Text, _>(bot_name)
+        .get_result(&mut conn)
+        .map_err(|e| format!("Failed to check if bot exists: {e}"))?;
+
+        Ok(exists.exists)
+    }
+
     #[cfg(feature = "llm")]
     pub async fn stream_response(
         &self,
@@ -90,6 +219,7 @@ impl BotOrchestrator {
 
         let user_id = Uuid::parse_str(&message.user_id)?;
         let session_id = Uuid::parse_str(&message.session_id)?;
+        let message_content = message.content.clone();
 
         let (session, context_data, history, model, key) = {
             let state_clone = self.state.clone();
@@ -117,13 +247,24 @@ impl BotOrchestrator {
                     };
 
                     let config_manager = ConfigManager::new(state_clone.conn.clone());
+
+                    // DEBUG: Log which bot we're getting config for
+                    info!("[CONFIG_TRACE] Getting LLM config for bot_id: {}", session.bot_id);
+
                     let model = config_manager
                         .get_config(&session.bot_id, "llm-model", Some("gpt-3.5-turbo"))
                         .unwrap_or_else(|_| "gpt-3.5-turbo".to_string());
+
                     let key = config_manager
                         .get_config(&session.bot_id, "llm-key", Some(""))
                         .unwrap_or_default();
 
+                    // DEBUG: Log the exact config values retrieved
+                    info!("[CONFIG_TRACE] Model: '{}'", model);
+                    info!("[CONFIG_TRACE] API Key: '{}' ({} chars)", key, key.len());
+                    info!("[CONFIG_TRACE] API Key first 10 chars: '{}'", &key.chars().take(10).collect::<String>());
+                    info!("[CONFIG_TRACE] API Key last 10 chars: '{}'", &key.chars().rev().take(10).collect::<String>());
+
                     Ok((session, context_data, history, model, key))
                 },
             )
@@ -131,7 +272,39 @@ impl BotOrchestrator {
         };
 
         let system_prompt = "You are a helpful assistant.".to_string();
-        let messages = OpenAIClient::build_messages(&system_prompt, &context_data, &history);
+        let mut messages = OpenAIClient::build_messages(&system_prompt, &context_data, &history);
+
+        #[cfg(any(feature = "research", feature = "llm"))]
+        {
+            if let Some(kb_manager) = self.state.kb_manager.as_ref() {
+                let bot_name_for_kb = {
+                    let conn = self.state.conn.get().ok();
+                    if let Some(mut db_conn) = conn {
+                        use crate::shared::models::schema::bots::dsl::*;
+                        bots.filter(id.eq(session.bot_id))
+                            .select(name)
+                            .first::<String>(&mut db_conn)
+                            .unwrap_or_else(|_| "default".to_string())
+                    } else {
+                        "default".to_string()
+                    }
+                };
+
+                if let Err(e) = inject_kb_context(
+                    kb_manager.clone(),
+                    self.state.conn.clone(),
+                    session_id,
+                    &bot_name_for_kb,
+                    &message_content,
+                    &mut messages,
+                    8000,
+                )
+                .await
+                {
+                    error!("Failed to inject KB context: {}", e);
+                }
+            }
+        }
 
         let (stream_tx, mut stream_rx) = mpsc::channel::<String>(100);
         let llm = self.state.llm_provider.clone();
@@ -139,6 +312,16 @@ impl BotOrchestrator {
         let model_clone = model.clone();
         let key_clone = key.clone();
         let messages_clone = messages.clone();
+
+        // DEBUG: Log exact values being passed to LLM
+        info!("[LLM_CALL] Calling generate_stream with:");
+        info!("[LLM_CALL]   Model: '{}'", model_clone);
+        info!("[LLM_CALL]   Key length: {} chars", key_clone.len());
+        info!("[LLM_CALL]   Key preview: '{}...{}'",
+            &key_clone.chars().take(8).collect::<String>(),
+            &key_clone.chars().rev().take(8).collect::<String>()
+        );
+
         tokio::spawn(async move {
             if let Err(e) = llm
                 .generate_stream("", &messages_clone, stream_tx, &model_clone, &key_clone)
@@ -388,10 +571,18 @@ pub async fn websocket_handler(
         let conn = state.conn.get().ok();
         if let Some(mut db_conn) = conn {
             use crate::shared::models::schema::bots::dsl::*;
-            let result: Result<Uuid, _> = bots
-                .filter(name.eq(&bot_name))
-                .select(id)
-                .first(&mut db_conn);
+
+            // Try to parse as UUID first, if that fails treat as bot name
+            let result: Result<Uuid, _> = if let Ok(uuid) = Uuid::parse_str(&bot_name) {
+                // Parameter is a UUID, look up by id
+                bots.filter(id.eq(uuid)).select(id).first(&mut db_conn)
+            } else {
+                // Parameter is a bot name, look up by name
+                bots.filter(name.eq(&bot_name))
+                    .select(id)
+                    .first(&mut db_conn)
+            };
+
             result.unwrap_or_else(|_| {
                 log::warn!("Bot not found: {}, using nil bot_id", bot_name);
                 Uuid::nil()
@@ -427,8 +618,8 @@ async fn handle_websocket(
     }
 
     info!(
-        "WebSocket connected for session: {}, user: {}",
-        session_id, user_id
+        "WebSocket connected for session: {}, user: {}, bot: {}",
+        session_id, user_id, bot_id
     );
 
     let welcome = serde_json::json!({
@@ -445,6 +636,89 @@ async fn handle_websocket(
         }
     }
 
+    // Execute start.bas automatically on connection (similar to auth.ast pattern)
+    {
+        let bot_name_result = {
+            let conn = state.conn.get().ok();
+            if let Some(mut db_conn) = conn {
+                use crate::shared::models::schema::bots::dsl::*;
+                bots.filter(id.eq(bot_id))
+                    .select(name)
+                    .first::<String>(&mut db_conn)
+                    .ok()
+            } else {
+                None
+            }
+        };
+
+        // DEBUG: Log start script execution attempt
+        info!(
+            "Checking for start.bas: bot_id={}, bot_name_result={:?}",
+            bot_id,
+            bot_name_result
+        );
+
+        if let Some(bot_name) = bot_name_result {
+            let start_script_path = format!("./work/{}.gbai/{}.gbdialog/start.bas", bot_name, bot_name);
+
+            info!("Looking for start.bas at: {}", start_script_path);
+
+            if let Ok(metadata) = tokio::fs::metadata(&start_script_path).await {
+                if metadata.is_file() {
+                    info!("Found start.bas file, reading contents...");
+                    if let Ok(start_script) = tokio::fs::read_to_string(&start_script_path).await {
+                        info!(
+                            "Executing start.bas for bot {} on session {}",
+                            bot_name, session_id
+                        );
+
+                        let state_for_start = state.clone();
+                        let _tx_for_start = tx.clone();
+
+                        tokio::spawn(async move {
+                            let session_result = {
+                                let mut sm = state_for_start.session_manager.lock().await;
+                                sm.get_session_by_id(session_id)
+                            };
+
+                            if let Ok(Some(session)) = session_result {
+                                info!("Executing start.bas for bot {} on session {}", bot_name, session_id);
+
+                                let result = tokio::task::spawn_blocking(move || {
+                                    let mut script_service = crate::basic::ScriptService::new(
+                                        state_for_start.clone(),
+                                        session.clone()
+                                    );
+                                    script_service.load_bot_config_params(&state_for_start, bot_id);
+
+                                    match script_service.compile(&start_script) {
+                                        Ok(ast) => match script_service.run(&ast) {
+                                            Ok(_) => Ok(()),
+                                            Err(e) => Err(format!("Script execution error: {}", e)),
+                                        },
+                                        Err(e) => Err(format!("Script compilation error: {}", e)),
+                                    }
+                                }).await;
+
+                                match result {
+                                    Ok(Ok(())) => {
+                                        info!("start.bas executed successfully for bot {}", bot_name);
+                                    }
+                                    Ok(Err(e)) => {
+                                        error!("start.bas error for bot {}: {}", bot_name, e);
+                                    }
+                                    Err(e) => {
+                                        error!("start.bas task error for bot {}: {}", bot_name, e);
+                                    }
+                                }
+                            }
+                        });
+                    }
+                }
+            }
+        }
+    }
+
     let mut send_task = tokio::spawn(async move {
         while let Some(response) = rx.recv().await {
             if let Ok(json_str) = serde_json::to_string(&response) {

src/core/config/mod.rs

@@ -362,14 +362,55 @@ impl ConfigManager {
         use crate::shared::models::schema::bot_configuration::dsl::*;
         let mut conn = self.get_conn()?;
         let fallback_str = fallback.unwrap_or("");
+
+        // Helper function to check if a value should be treated as "not configured"
+        fn is_placeholder_value(value: &str) -> bool {
+            let trimmed = value.trim().to_lowercase();
+            trimmed.is_empty() || trimmed == "none" || trimmed == "null" || trimmed == "n/a"
+        }
+
+        // Helper function to check if a value is a local file path (for local LLM server)
+        // These should fall back to default bot's config when using remote API
+        fn is_local_file_path(value: &str) -> bool {
+            let value = value.trim();
+            // Check for file path patterns
+            value.starts_with("../") ||
+            value.starts_with("./") ||
+            value.starts_with('/') ||
+            value.starts_with("~") ||
+            value.contains(".gguf") ||
+            value.contains(".bin") ||
+            value.contains(".safetensors") ||
+            value.starts_with("data/") ||
+            value.starts_with("../../") ||
+            value.starts_with("models/")
+        }
+
+        // Try to get value for the specific bot
         let result = bot_configuration
             .filter(bot_id.eq(code_bot_id))
             .filter(config_key.eq(key))
             .select(config_value)
             .first::<String>(&mut conn);
+
         let value = match result {
-            Ok(v) => v,
+            Ok(v) => {
+                // Check if it's a placeholder value or local file path - if so, fall back to default bot
+                // Local file paths are valid for local LLM server but NOT for remote APIs
+                if is_placeholder_value(&v) || is_local_file_path(&v) {
+                    let (default_bot_id, _default_bot_name) = crate::bot::get_default_bot(&mut conn);
+                    bot_configuration
+                        .filter(bot_id.eq(default_bot_id))
+                        .filter(config_key.eq(key))
+                        .select(config_value)
+                        .first::<String>(&mut conn)
+                        .unwrap_or_else(|_| fallback_str.to_string())
+                } else {
+                    v
+                }
+            }
             Err(_) => {
+                // Value not found, fall back to default bot
                 let (default_bot_id, _default_bot_name) = crate::bot::get_default_bot(&mut conn);
                 bot_configuration
                     .filter(bot_id.eq(default_bot_id))
@@ -379,7 +420,15 @@ impl ConfigManager {
                     .unwrap_or_else(|_| fallback_str.to_string())
             }
         };
-        Ok(value)
+
+        // Final check: if the result is still a placeholder value, use the fallback_str
+        let final_value = if is_placeholder_value(&value) {
+            fallback_str.to_string()
+        } else {
+            value
+        };
+
+        Ok(final_value)
     }
 
     pub fn get_bot_config_value(

src/llm/mod.rs

@@ -1,6 +1,6 @@
 use async_trait::async_trait;
 use futures::StreamExt;
-use log::{info, trace};
+use log::{error, info};
 use serde_json::Value;
 use std::sync::Arc;
 use tokio::sync::{mpsc, RwLock};
@@ -44,13 +44,116 @@ pub trait LLMProvider: Send + Sync {
 pub struct OpenAIClient {
     client: reqwest::Client,
     base_url: String,
+    endpoint_path: String,
 }
 
 impl OpenAIClient {
-    pub fn new(_api_key: String, base_url: Option<String>) -> Self {
+    /// Estimates token count for a text string (roughly 4 characters per token for English)
+    fn estimate_tokens(text: &str) -> usize {
+        // Rough estimate: ~4 characters per token for English text
+        // This is a heuristic and may not be accurate for all languages
+        text.len().div_ceil(4)
+    }
+
+    /// Estimates total tokens for a messages array
+    fn estimate_messages_tokens(messages: &Value) -> usize {
+        if let Some(msg_array) = messages.as_array() {
+            msg_array
+                .iter()
+                .map(|msg| {
+                    if let Some(content) = msg.get("content").and_then(|c| c.as_str()) {
+                        Self::estimate_tokens(content)
+                    } else {
+                        0
+                    }
+                })
+                .sum()
+        } else {
+            0
+        }
+    }
+
+    /// Truncates messages to fit within the max_tokens limit
+    /// Keeps system messages and the most recent user/assistant messages
+    fn truncate_messages(messages: &Value, max_tokens: usize) -> Value {
+        let mut result = Vec::new();
+        let mut token_count = 0;
+
+        if let Some(msg_array) = messages.as_array() {
+            // First pass: keep all system messages
+            for msg in msg_array {
+                if let Some(role) = msg.get("role").and_then(|r| r.as_str()) {
+                    if role == "system" {
+                        if let Some(content) = msg.get("content").and_then(|c| c.as_str()) {
+                            let msg_tokens = Self::estimate_tokens(content);
+                            if token_count + msg_tokens <= max_tokens {
+                                result.push(msg.clone());
+                                token_count += msg_tokens;
+                            }
+                        }
+                    }
+                }
+            }
+
+            // Second pass: add user/assistant messages from newest to oldest
+            let mut recent_messages: Vec<&Value> = msg_array
+                .iter()
+                .filter(|msg| msg.get("role").and_then(|r| r.as_str()) != Some("system"))
+                .collect();
+
+            // Reverse to get newest first
+            recent_messages.reverse();
+
+            for msg in recent_messages {
+                if let Some(content) = msg.get("content").and_then(|c| c.as_str()) {
+                    let msg_tokens = Self::estimate_tokens(content);
+                    if token_count + msg_tokens <= max_tokens {
+                        result.push(msg.clone());
+                        token_count += msg_tokens;
+                    } else {
+                        break;
+                    }
+                }
+            }
+
+            // Reverse back to chronological order for non-system messages
+            // But keep system messages at the beginning
+            let system_count = result.len()
+                - result
+                    .iter()
+                    .filter(|m| m.get("role").and_then(|r| r.as_str()) != Some("system"))
+                    .count();
+            let mut user_messages: Vec<Value> = result.drain(system_count..).collect();
+            user_messages.reverse();
+            result.extend(user_messages);
+        }
+
+        serde_json::Value::Array(result)
+    }
+
+    /// Ensures messages fit within model's context limit
+    fn ensure_token_limit(messages: &Value, model_context_limit: usize) -> Value {
+        let estimated_tokens = Self::estimate_messages_tokens(messages);
+
+        // Use 90% of context limit to leave room for response
+        let safe_limit = (model_context_limit as f64 * 0.9) as usize;
+
+        if estimated_tokens > safe_limit {
+            log::warn!(
+                "Messages exceed token limit ({} > {}), truncating...",
+                estimated_tokens,
+                safe_limit
+            );
+            Self::truncate_messages(messages, safe_limit)
+        } else {
+            messages.clone()
+        }
+    }
+    pub fn new(_api_key: String, base_url: Option<String>, endpoint_path: Option<String>) -> Self {
         Self {
             client: reqwest::Client::new(),
             base_url: base_url.unwrap_or_else(|| "https://api.openai.com".to_string()),
+            endpoint_path: endpoint_path.unwrap_or_else(|| "/v1/chat/completions".to_string()),
         }
     }
 
@@ -92,21 +195,64 @@ impl LLMProvider for OpenAIClient {
         key: &str,
     ) -> Result<String, Box<dyn std::error::Error + Send + Sync>> {
         let default_messages = serde_json::json!([{"role": "user", "content": prompt}]);
+
+        // Get the messages to use
+        let raw_messages =
+            if messages.is_array() && !messages.as_array().unwrap_or(&vec![]).is_empty() {
+                messages
+            } else {
+                &default_messages
+            };
+
+        // Ensure messages fit within model's context limit
+        // GLM-4.7 has 202750 tokens, other models vary
+        let context_limit = if model.contains("glm-4") || model.contains("GLM-4") {
+            202750
+        } else if model.contains("gpt-4") {
+            128000
+        } else if model.contains("gpt-3.5") {
+            16385
+        } else if model.starts_with("http://localhost:808") || model == "local" {
+            768 // Local llama.cpp server context limit
+        } else {
+            4096 // Default conservative limit
+        };
+
+        let messages = OpenAIClient::ensure_token_limit(raw_messages, context_limit);
+
+        let full_url = format!("{}{}", self.base_url, self.endpoint_path);
+        let auth_header = format!("Bearer {}", key);
+
+        // Debug logging to help troubleshoot 401 errors
+        info!("LLM Request Details:");
+        info!("  URL: {}", full_url);
+        info!("  Authorization: Bearer <{} chars>", key.len());
+        info!("  Model: {}", model);
+        if let Some(msg_array) = messages.as_array() {
+            info!("  Messages: {} messages", msg_array.len());
+        }
+        info!("  API Key First 8 chars: '{}...'", &key.chars().take(8).collect::<String>());
+        info!("  API Key Last 8 chars: '...{}'", &key.chars().rev().take(8).collect::<String>());
+
         let response = self
             .client
-            .post(format!("{}/v1/chat/completions", self.base_url))
-            .header("Authorization", format!("Bearer {}", key))
+            .post(&full_url)
+            .header("Authorization", &auth_header)
             .json(&serde_json::json!({
                 "model": model,
-                "messages": if messages.is_array() && !messages.as_array().unwrap_or(&vec![]).is_empty() {
-                    messages
-                } else {
-                    &default_messages
-                }
+                "messages": messages,
+                "stream": true
             }))
             .send()
             .await?;
 
+        let status = response.status();
+        if status != reqwest::StatusCode::OK {
+            let error_text = response.text().await.unwrap_or_default();
+            error!("LLM generate error: {}", error_text);
+            return Err(format!("LLM request failed with status: {}", status).into());
+        }
+
         let result: Value = response.json().await?;
         let raw_content = result["choices"][0]["message"]["content"]
             .as_str()
@@ -127,18 +273,51 @@ impl LLMProvider for OpenAIClient {
         key: &str,
     ) -> Result<(), Box<dyn std::error::Error + Send + Sync>> {
         let default_messages = serde_json::json!([{"role": "user", "content": prompt}]);
+
+        // Get the messages to use
+        let raw_messages =
+            if messages.is_array() && !messages.as_array().unwrap_or(&vec![]).is_empty() {
+                info!("Using provided messages: {:?}", messages);
+                messages
+            } else {
+                &default_messages
+            };
+
+        // Ensure messages fit within model's context limit
+        // GLM-4.7 has 202750 tokens, other models vary
+        let context_limit = if model.contains("glm-4") || model.contains("GLM-4") {
+            202750
+        } else if model.contains("gpt-4") {
+            128000
+        } else if model.contains("gpt-3.5") {
+            16385
+        } else if model.starts_with("http://localhost:808") || model == "local" {
+            768 // Local llama.cpp server context limit
+        } else {
+            4096 // Default conservative limit
+        };
+
+        let messages = OpenAIClient::ensure_token_limit(raw_messages, context_limit);
+
+        let full_url = format!("{}{}", self.base_url, self.endpoint_path);
+        let auth_header = format!("Bearer {}", key);
+
+        // Debug logging to help troubleshoot 401 errors
+        info!("LLM Request Details:");
+        info!("  URL: {}", full_url);
+        info!("  Authorization: Bearer <{} chars>", key.len());
+        info!("  Model: {}", model);
+        if let Some(msg_array) = messages.as_array() {
+            info!("  Messages: {} messages", msg_array.len());
+        }
+
         let response = self
             .client
-            .post(format!("{}/v1/chat/completions", self.base_url))
-            .header("Authorization", format!("Bearer {}", key))
+            .post(&full_url)
+            .header("Authorization", &auth_header)
             .json(&serde_json::json!({
                 "model": model,
-                "messages": if messages.is_array() && !messages.as_array().unwrap_or(&vec![]).is_empty() {
-                    info!("Using provided messages: {:?}", messages);
-                    messages
-                } else {
-                    &default_messages
-                },
+                "messages": messages,
                 "stream": true
             }))
             .send()
@@ -147,7 +326,7 @@ impl LLMProvider for OpenAIClient {
         let status = response.status();
         if status != reqwest::StatusCode::OK {
             let error_text = response.text().await.unwrap_or_default();
-            trace!("LLM generate_stream error: {}", error_text);
+            error!("LLM generate_stream error: {}", error_text);
             return Err(format!("LLM request failed with status: {}", status).into());
         }
 
@@ -213,11 +392,16 @@ pub fn create_llm_provider(
     provider_type: LLMProviderType,
     base_url: String,
     deployment_name: Option<String>,
+    endpoint_path: Option<String>,
 ) -> std::sync::Arc<dyn LLMProvider> {
     match provider_type {
         LLMProviderType::OpenAI => {
             info!("Creating OpenAI LLM provider with URL: {}", base_url);
-            std::sync::Arc::new(OpenAIClient::new("empty".to_string(), Some(base_url)))
+            std::sync::Arc::new(OpenAIClient::new(
+                "empty".to_string(),
+                Some(base_url),
+                endpoint_path,
+            ))
         }
         LLMProviderType::Claude => {
             info!("Creating Claude LLM provider with URL: {}", base_url);
@@ -237,9 +421,10 @@ pub fn create_llm_provider(
 pub fn create_llm_provider_from_url(
     url: &str,
     model: Option<String>,
+    endpoint_path: Option<String>,
 ) -> std::sync::Arc<dyn LLMProvider> {
     let provider_type = LLMProviderType::from(url);
-    create_llm_provider(provider_type, url.to_string(), model)
+    create_llm_provider(provider_type, url.to_string(), model, endpoint_path)
 }
 
 pub struct DynamicLLMProvider {
@@ -259,8 +444,13 @@ impl DynamicLLMProvider {
         info!("LLM provider updated dynamically");
     }
 
-    pub async fn update_from_config(&self, url: &str, model: Option<String>) {
-        let new_provider = create_llm_provider_from_url(url, model);
+    pub async fn update_from_config(
+        &self,
+        url: &str,
+        model: Option<String>,
+        endpoint_path: Option<String>,
+    ) {
+        let new_provider = create_llm_provider_from_url(url, model, endpoint_path);
         self.update_provider(new_provider).await;
     }
 
@@ -490,7 +680,7 @@ mod tests {
 
     #[test]
     fn test_openai_client_new_default_url() {
-        let client = OpenAIClient::new("test_key".to_string(), None);
+        let client = OpenAIClient::new("test_key".to_string(), None, None);
         assert_eq!(client.base_url, "https://api.openai.com");
     }
 
@@ -499,6 +689,7 @@ mod tests {
         let client = OpenAIClient::new(
             "test_key".to_string(),
             Some("http://localhost:8080".to_string()),
+            None,
         );
         assert_eq!(client.base_url, "http://localhost:8080");
     }