diff --git a/src/index.ts b/src/index.ts
index 4b912cd..d186e5a 100644
--- a/src/index.ts
+++ b/src/index.ts
@@ -19,10 +19,15 @@ import * as fs from "fs";
 import * as path from "path";
 
 // Configuration - can be overridden via environment variables
-const API_BASE_URL = process.env.SECOND_BRAIN_API_URL || "https://2brain.coer.nl/api";
+// API_BASE_URL should point to the base (e.g., http://192.168.1.16:8001)
+// The /api/v1 prefix is added by apiRequest
+const API_BASE_URL = (process.env.SECOND_BRAIN_API_URL || "http://192.168.1.16:8001").replace(/\/+$/, "");
 const API_USERNAME = process.env.SECOND_BRAIN_USERNAME || "";
 const API_PASSWORD = process.env.SECOND_BRAIN_PASSWORD || "";
 
+// Default search threshold - embedding scores are often low, so use a low default
+const DEFAULT_SEARCH_THRESHOLD = 0.01;
+
 // Build auth header if credentials provided
 function getAuthHeader(): Record<string, string> {
   if (API_USERNAME && API_PASSWORD) {
@@ -37,7 +42,11 @@ async function apiRequest(
   endpoint: string,
   options: RequestInit = {}
 ): Promise<Response> {
-  const url = `${API_BASE_URL}${endpoint}`;
+  // Ensure endpoint starts with /api/v1
+  const normalizedEndpoint = endpoint.startsWith("/api/v1") 
+    ? endpoint 
+    : `/api${endpoint.startsWith("/v1") ? endpoint : `/v1${endpoint}`}`;
+  const url = `${API_BASE_URL}${normalizedEndpoint}`;
   const headers = {
     ...getAuthHeader(),
     ...options.headers,
@@ -321,12 +330,17 @@ async function handleSearch(args: {
   query: string;
   top_k?: number;
 }): Promise<string> {
-  const params = new URLSearchParams({
-    q: args.query,
-    top_k: String(args.top_k || 5),
+  // Use POST with JSON body for more control over search parameters
+  const response = await apiRequest(`/v1/search`, {
+    method: "POST",
+    headers: { "Content-Type": "application/json" },
+    body: JSON.stringify({
+      query: args.query,
+      limit: args.top_k || 5,
+      threshold: DEFAULT_SEARCH_THRESHOLD,
+      hybrid: true,
+    }),
   });
-
-  const response = await apiRequest(`/v1/search?${params}`);
   const data = await response.json();
 
   if (!response.ok) {
@@ -450,31 +464,35 @@ async function handleListDocuments(args: {
   folder?: string;
   limit?: number;
 }): Promise<string> {
-  const params = new URLSearchParams({
-    limit: String(args.limit || 20),
-  });
-  if (args.folder) params.append("folder", args.folder);
-
-  const response = await apiRequest(`/v1/documents?${params}`);
-  const data = await response.json();
+  // The /documents endpoint doesn't exist, so we use search with a broad query
+  // or get stats to show document count
+  const statsResponse = await apiRequest("/v1/stats");
+  const stats = await statsResponse.json();
 
-  if (!response.ok) {
+  if (!statsResponse.ok) {
     throw new Error(
-      `Failed to list documents: ${data.detail || response.statusText}`
+      `Failed to get stats: ${stats.detail || statsResponse.statusText}`
     );
   }
 
-  if (!data.documents || data.documents.length === 0) {
-    return "No documents found.";
-  }
+  // Also get tags to show what's in the vault
+  const tagsResponse = await apiRequest("/v1/tags");
+  const tags = await tagsResponse.json();
+
+  let output = `📚 Knowledge Base Overview:\n\n`;
+  output += `- **Documents:** ${stats.total_documents || 0}\n`;
+  output += `- **Chunks:** ${stats.total_chunks || 0}\n`;
+  output += `- **Last indexed:** ${stats.last_indexed || "Never"}\n\n`;
 
-  let output = `📚 Documents (${data.total}):\n\n`;
-  for (const doc of data.documents) {
-    output += `- **${doc.title || doc.path}**`;
-    if (doc.tags?.length) output += ` [${doc.tags.join(", ")}]`;
-    output += "\n";
+  if (Array.isArray(tags) && tags.length > 0) {
+    output += `**Tags:**\n`;
+    for (const tag of tags) {
+      output += `- #${tag.tag} (${tag.count} docs)\n`;
+    }
   }
 
+  output += `\n_Use search to find specific documents._`;
+
   return output;
 }
 
@@ -489,10 +507,12 @@ async function handleGetStats(): Promise<string> {
   }
 
   return `📊 Second Brain Stats:
-- Documents: ${data.document_count || 0}
-- Chunks: ${data.chunk_count || 0}
-- Vault size: ${data.vault_size_mb?.toFixed(2) || 0} MB
-- Last indexed: ${data.last_indexed || "Never"}`;
+- Documents: ${data.total_documents || data.document_count || 0}
+- Chunks: ${data.total_chunks || data.chunk_count || 0}
+- Vault size: ${data.vault_size_mb?.toFixed(2) || "N/A"} MB
+- Last indexed: ${data.last_indexed || "Never"}
+- Embedding model: ${data.embedding_model || "N/A"}
+- Chat model: ${data.chat_model || "N/A"}`;
 }
 
 async function handleReindex(): Promise<string> {