@laskarks/mcp-rag-node 1.0.6 → 1.0.7

package/README.md

@@ -50,7 +50,7 @@ For OpenRouter, use the model ID format, e.g. `openai/text-embedding-3-small` or
 
 | Variable               | Description                           | Default |
 | ---------------------- | ------------------------------------- | ------- |
-| `RAG_CHUNK_MAX_TOKENS` | Max tokens per chunk before embedding | `512`   |
+| `RAG_CHUNK_MAX_TOKENS` | Max tokens per chunk before embedding | `1536`  |
 | `RAG_CHUNK_OVERLAP`    | Overlap tokens between chunks         | `50`    |
 
 ## Usage
@@ -88,10 +88,11 @@ Install the package first: `npm i @laskarks/mcp-rag-node`
   "mcpServers": {
     "rag": {
       "command": "npx",
-      "args": ["-y", "@laskarks/mcp-rag-node"],
+      "args": ["-y", "@laskarks/mcp-rag-node@latest"],
       "env": {
         "APIKEY": "sk-...",
         "EMBEDDING_MODEL": "text-embedding-3-small",
+        "RAG_CHUNK_MAX_TOKENS": 1536,
         "PINECONE_API_KEY": "...",
         "PINECONE_INDEX": "rag-index",
         "PROVIDER": "openai"
@@ -108,10 +109,11 @@ Install the package first: `npm i @laskarks/mcp-rag-node`
   "mcpServers": {
     "rag": {
       "command": "npx",
-      "args": ["-y", "@laskarks/mcp-rag-node"],
+      "args": ["-y", "@laskarks/mcp-rag-node@latest"],
       "env": {
         "APIKEY": "sk-...",
         "EMBEDDING_MODEL": "text-embedding-3-small",
+        "RAG_CHUNK_MAX_TOKENS": 1536,
         "PINECONE_API_KEY": "...",
         "PINECONE_INDEX": "rag-index",
         "PROVIDER": "openai"

package/dist/ai.js

@@ -52,7 +52,7 @@ class AI {
     chunkByToken(text, maxTokens, overlap) {
         const defaultMax = env.RAG_CHUNK_MAX_TOKENS
             ? Number(env.RAG_CHUNK_MAX_TOKENS)
-            : 512;
+            : 1536;
         const defaultOverlap = 50;
         const limit = maxTokens ?? defaultMax;
         const overlapTokens = overlap ?? defaultOverlap;
@@ -129,13 +129,13 @@ class AI {
                     topK: 3,
                     includeMetadata: true,
                 });
-                return results;
-                // const relevantChunks = results.matches.map((match) => ({
-                //   text: match.metadata?.text as string,
-                //   score: match.score,
-                // }));
-                // const context = relevantChunks.map((c) => c.text).join("\n\n");
-                // return context;
+                // Get text from metadata
+                const relevantChunks = results.matches.map((match) => ({
+                    text: match.metadata?.text,
+                    score: match.score,
+                }));
+                const context = relevantChunks.map((c) => c.text).join("\n\n");
+                return context;
             }
             else {
                 return response?.data[0] || "Unexpected error";

package/dist/index.js

@@ -60,7 +60,7 @@ async function main() {
     }, async ({ keyword }) => {
         const response = await CallAI.search_documents(keyword);
         return {
-            content: [{ type: "text", text: `${response}` }],
+            content: [{ type: "text", text: JSON.stringify(response, null, 2) }],
         };
     });
     const transport = new StdioServerTransport();

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@laskarks/mcp-rag-node",
-  "version": "1.0.6",
+  "version": "1.0.7",
   "description": "Simple MCP RAG server using @modelcontextprotocol/sdk",
   "main": "dist/index.js",
   "bin": {

package/src/ai.ts

@@ -68,7 +68,7 @@ class AI {
   chunkByToken(text: string, maxTokens?: number, overlap?: number): string[] {
     const defaultMax = env.RAG_CHUNK_MAX_TOKENS
       ? Number(env.RAG_CHUNK_MAX_TOKENS)
-      : 512;
+      : 1536;
     const defaultOverlap = 50;
     const limit = maxTokens ?? defaultMax;
     const overlapTokens = overlap ?? defaultOverlap;