prism-mcp-server 7.8.1 → 7.8.3

package/README.md

@@ -691,7 +691,7 @@ The Generator strips the `console.log`, resubmits, and the next `EVALUATE` retur
 
 ## 🆕 What's New
 
-> **Current release: v7.8.0 — Cognitive Architecture**
+> **Current release: v7.8.2 — Cognitive Architecture**
 
 - 🧠 **v7.8.0 — Cognitive Architecture:** The biggest leap forward yet. Moved beyond flat vector search into a true cognitive architecture inspired by human brain mechanics. Episodic-to-Semantic memory consolidation (Hebbian learning), ACT-R Spreading Activation with multi-hop causal reasoning, Uncertainty-Aware Rejection Gate (your agent can say "I don't know"), and Dynamic Fast Weight Decay (semantic memories outlive episodic chatter by 2×). **Your agents don't just remember; they learn.** → [Cognitive Architecture](#-cognitive-architecture-v78)
 - 🌐 **v7.7.0 — Cloud-Native SSE Transport:** Full unauthenticated and authenticated Server-Sent Events MCP support for seamless network deployments.
@@ -1065,7 +1065,7 @@ Prism has evolved from smart session logging into a **cognitive memory architect
 
 ## 📦 Milestones & Roadmap
 
-> **Current: v7.8.0** — Cognitive Architecture ([CHANGELOG](CHANGELOG.md))
+> **Current: v7.8.2** — Cognitive Architecture ([CHANGELOG](CHANGELOG.md))
 
 | Release | Headline |
 |---------|----------|

package/dist/utils/llm/adapters/voyage.js

@@ -52,9 +52,9 @@ const EMBEDDING_DIMS = 768;
 // with OpenAI and Gemini adapters) avoids tokenizer dependency.
 // 8000 chars ≈ 1500-2000 tokens for typical session summaries.
 const MAX_EMBEDDING_CHARS = 8000;
-// Default model: voyage-3 (supports output_dimension=768 via MRL)
-// voyage-3-lite is NOT recommended as its native 512 dims < 768.
-const DEFAULT_MODEL = "voyage-3";
+// Default model: voyage-code-3 (supports output_dimension=768 via client-side MRL truncation)
+// Extremely optimized for code bases, ide workspaces, and technical queries.
+const DEFAULT_MODEL = "voyage-code-3";
 const VOYAGE_API_BASE = "https://api.voyageai.com/v1";
 // ─── Adapter ─────────────────────────────────────────────────────────────────
 export class VoyageAdapter {
@@ -91,11 +91,9 @@ export class VoyageAdapter {
         const requestBody = {
             input: [truncated],
             model,
-            // Request exactly 768 dims via Matryoshka truncation.
-            // Supported by voyage-3, voyage-3-large, voyage-code-3.
-            // voyage-3-lite (native 512 dims) will ignore this and return 512,
-            // which will be caught by the dimension guard below.
-            output_dimension: EMBEDDING_DIMS,
+            // We do NOT send output_dimension here because Voyage's API explicitly
+            // restricts it to [256, 512, 1024, 2048] for MRL models. We will
+            // manually slice the 1024-dim result down to 768 client-side.
         };
         const response = await fetch(`${VOYAGE_API_BASE}/embeddings`, {
             method: "POST",
@@ -110,17 +108,20 @@ export class VoyageAdapter {
             throw new Error(`[VoyageAdapter] API request failed — status=${response.status}: ${errorText}`);
         }
         const data = (await response.json());
-        const embedding = data?.data?.[0]?.embedding;
+        let embedding = data?.data?.[0]?.embedding;
         if (!Array.isArray(embedding)) {
             throw new Error("[VoyageAdapter] Unexpected response format — no embedding array found");
         }
+        // Client-side MRL Truncation: 
+        // Voyage models returning 1024 dims can be safely sliced to 768 since they 
+        // are trained with Matryoshka Representation Learning.
+        if (embedding.length > EMBEDDING_DIMS) {
+            embedding = embedding.slice(0, EMBEDDING_DIMS);
+        }
         // Dimension guard: Prism's DB schema requires exactly 768 dims.
-        // This catches voyage-3-lite (512) or future API changes silently early.
         if (embedding.length !== EMBEDDING_DIMS) {
             throw new Error(`[VoyageAdapter] Embedding dimension mismatch: expected ${EMBEDDING_DIMS}, ` +
-                `got ${embedding.length}. ` +
-                `Use voyage-3 (not voyage-3-lite) to get 768-dim output via MRL truncation. ` +
-                `Change voyage_model in the Mind Palace dashboard.`);
+                `got ${embedding.length}. Make sure you are using a model that returns at least 768 dims.`);
         }
         debugLog(`[VoyageAdapter] Embedding generated — dims=${embedding.length}, ` +
             `tokens_used=${data.usage?.total_tokens ?? "unknown"}`);

package/dist/utils/llm/factory.js

@@ -89,9 +89,16 @@ export function getLLMProvider() {
     // ── Resolve embedding provider ────────────────────────────────────────
     let embedType = getSettingSync("embedding_provider", "auto");
     if (embedType === "auto") {
-        // Anthropic has no embedding API — auto-bridge to Gemini.
-        // For all other text providers, use the same provider for embeddings.
-        embedType = textType === "anthropic" ? "gemini" : textType;
+        if (process.env.VOYAGE_API_KEY) {
+            // If Voyage is available, use it as the default embedding provider
+            // since voyage-code-3 strongly outperforms general embeddings on code contexts.
+            embedType = "voyage";
+        }
+        else {
+            // Anthropic has no embedding API — auto-bridge to Gemini.
+            // For all other text providers, use the same provider for embeddings.
+            embedType = textType === "anthropic" ? "gemini" : textType;
+        }
         if (textType === "anthropic") {
             console.info("[LLMFactory] text_provider=anthropic with embedding_provider=auto: " +
                 "routing embeddings to GeminiAdapter (Anthropic has no native embedding API). " +

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "prism-mcp-server",
-  "version": "7.8.1",
+  "version": "7.8.3",
   "mcpName": "io.github.dcostenco/prism-mcp",
   "description": "The Mind Palace for AI Agents — a true Cognitive Architecture with Hebbian learning (episodic→semantic consolidation), ACT-R spreading activation (multi-hop causal reasoning), uncertainty-aware rejection gates (agents that know when they don't know), adversarial evaluation (anti-sycophancy), fail-closed Dark Factory pipelines, persistent memory (SQLite/Supabase), multi-agent Hivemind, time travel & visual dashboard. Zero-config local mode.",
   "module": "index.ts",