prism-mcp-server 2.1.0 → 2.1.1

package/README.md

@@ -44,17 +44,15 @@ Add this to your `claude_desktop_config.json` or `.cursor/mcp.json`:
   "mcpServers": {
     "prism-mcp": {
       "command": "npx",
-      "args": ["-y", "prism-mcp-server"],
-      "env": {
-        "PRISM_STORAGE": "local",
-        "BRAVE_API_KEY": "your-brave-api-key"
-      }
+      "args": ["-y", "prism-mcp-server"]
     }
   }
 }
 ```
 
-> **Note:** `BRAVE_API_KEY` is required for search features. To enable semantic search, Morning Briefings, and paper analysis, add `GOOGLE_API_KEY`. Everything else works out of the box with zero config.
+That's it — **zero env vars needed** for local memory, Mind Palace dashboard, Time Travel, and Telepathy.
+
+> **Optional API keys:** Add `BRAVE_API_KEY` for web search, `GOOGLE_API_KEY` for semantic search + Morning Briefings + paper analysis. See [Environment Variables](#environment-variables) for the full list.
 
 ### Option B: Cloud Sync Mode (Supabase)
 
@@ -68,7 +66,6 @@ To share memory across multiple machines or teams, switch to Supabase:
       "args": ["-y", "prism-mcp-server"],
       "env": {
         "PRISM_STORAGE": "supabase",
-        "BRAVE_API_KEY": "your-brave-api-key",
         "SUPABASE_URL": "https://your-project.supabase.co",
         "SUPABASE_KEY": "your-supabase-anon-key"
       }
@@ -95,7 +92,6 @@ Then add to your MCP config:
       "command": "node",
       "args": ["/absolute/path/to/prism-mcp/dist/server.js"],
       "env": {
-        "PRISM_STORAGE": "local",
         "BRAVE_API_KEY": "your-brave-api-key",
         "GOOGLE_API_KEY": "your-google-gemini-key"
       }
@@ -159,10 +155,7 @@ Add to your `claude_desktop_config.json`:
     "prism-mcp": {
       "command": "npx",
       "args": ["-y", "prism-mcp-server"],
-      "env": {
-        "PRISM_STORAGE": "local",
-        "BRAVE_API_KEY": "your-brave-api-key"
-      }
+      "env": {}
     }
   }
 }
@@ -181,10 +174,7 @@ Add to `.cursor/mcp.json` in your project root (or `~/.cursor/mcp.json` for glob
     "prism-mcp": {
       "command": "npx",
       "args": ["-y", "prism-mcp-server"],
-      "env": {
-        "PRISM_STORAGE": "local",
-        "BRAVE_API_KEY": "your-brave-api-key"
-      }
+      "env": {}
     }
   }
 }
@@ -203,10 +193,7 @@ Add to `~/.codeium/windsurf/mcp_config.json`:
     "prism-mcp": {
       "command": "npx",
       "args": ["-y", "prism-mcp-server"],
-      "env": {
-        "PRISM_STORAGE": "local",
-        "BRAVE_API_KEY": "your-brave-api-key"
-      }
+      "env": {}
     }
   }
 }
@@ -348,7 +335,7 @@ Instead of writing custom JavaScript, pass a `template` name for instant extract
 
 | Variable | Required | Description |
 |----------|----------|-------------|
-| `BRAVE_API_KEY` | Yes (for search) | Brave Search Pro API key |
+| `BRAVE_API_KEY` | No | Brave Search Pro API key (enables web/local search tools) |
 | `PRISM_STORAGE` | No | `"local"` (default) or `"supabase"` |
 | `GOOGLE_API_KEY` | No | Google AI / Gemini — enables paper analysis, Morning Briefings, compaction |
 | `BRAVE_ANSWERS_API_KEY` | No | Separate Brave Answers key for AI-grounded answers |

package/dist/utils/embeddingApi.js

@@ -1,10 +1,16 @@
 /**
- * Embedding Generation Utility (v0.4.0 — Enhancement #4)
+ * Embedding Generation Utility (v0.5.0 — Model Migration)
  *
  * ═══════════════════════════════════════════════════════════════════
- * REVIEWER NOTE: This module wraps Google's text-embedding-004 model
+ * REVIEWER NOTE: This module wraps Google's gemini-embedding-001 model
  * to generate 768-dimensional vector embeddings for text.
  *
+ * MIGRATION (2026-03): text-embedding-004 was deprecated Jan 14, 2026.
+ * Replaced with gemini-embedding-001 (GA on v1 API). The new model
+ * supports Matryoshka Representation Learning (MRL) for flexible
+ * output dimensions (768, 1536, 3072). We use 768 for compatibility
+ * with existing pgvector columns.
+ *
  * USAGE — Called in two places:
  *   1. sessionSaveLedgerHandler — embeds summary+decisions at save time
  *      (fire-and-forget, non-blocking)
@@ -16,25 +22,24 @@
  * Using a separate embedding service (OpenAI, Cohere) would add
  * another API key dependency and increase configuration complexity.
  *
- * COST: Gemini's text-embedding-004 is free tier for <1500 req/min.
+ * COST: Gemini's gemini-embedding-001 is free tier for <1500 req/min.
  * At typical usage (~10-50 ledger saves/day), we'll never approach
  * this limit.
  *
- * TRUNCATION GUARD: text-embedding-004 has a token limit per API call
- * (~8192 tokens ≈ ~32K characters). If the input text exceeds this,
- * the API returns a 400 Bad Request. We implement a hard character
- * limit (default 8000 chars) to guarantee the API call never crashes.
- * This is applied before sending to the API, not after.
+ * TRUNCATION GUARD: gemini-embedding-001 supports up to 2048 tokens
+ * per input. We implement a hard character limit (default 8000 chars)
+ * to guarantee the API call never crashes. This is applied before
+ * sending to the API, not after.
  * ═══════════════════════════════════════════════════════════════════
  */
 import { GoogleGenerativeAI } from "@google/generative-ai";
 import { GOOGLE_API_KEY } from "../config.js";
 // ─── Constants ────────────────────────────────────────────────
 // REVIEWER NOTE: Maximum characters to send to the embedding API.
-// text-embedding-004 supports ~8192 tokens. At ~4 chars per token,
+// gemini-embedding-001 supports up to 2048 tokens. At ~4 chars/token,
 // 8000 chars is a safe ceiling. Truncation is silent and non-fatal —
-// the embedding still captures the semantic meaning of the first
-// ~2000 tokens, which is more than enough for similarity search.
+// the embedding still captures the semantic meaning of the leading
+// content, which is more than enough for similarity search.
 const MAX_EMBEDDING_CHARS = 8000;
 // ─── Embedding Client ─────────────────────────────────────────
 /**
@@ -76,7 +81,8 @@ export async function generateEmbedding(text) {
         throw new Error("Cannot generate embedding for empty text");
     }
     const genAI = new GoogleGenerativeAI(GOOGLE_API_KEY);
-    const model = genAI.getGenerativeModel({ model: "text-embedding-004" });
+    const model = genAI.getGenerativeModel({ model: "gemini-embedding-001" }, { apiVersion: "v1beta" } // gemini-embedding-001 requires v1beta
+    );
     console.error(`[embedding] Generating 768-dim embedding for ${inputText.length} chars`);
     const result = await model.embedContent(inputText);
     return result.embedding.values;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "prism-mcp-server",
-  "version": "2.1.0",
+  "version": "2.1.1",
   "mcpName": "io.github.dcostenco/prism-mcp",
   "description": "The Mind Palace for AI Agents — local-first MCP server with persistent memory (SQLite/Supabase), visual dashboard, time travel, multi-agent sync, Morning Briefings, reality drift detection, code mode templates, semantic vector search, and Brave Search + Gemini analysis. Zero-config local mode.",
   "module": "index.ts",