mcpscraper-memory 1.22.0 → 1.23.0

package/CHANGELOG.md

@@ -4,6 +4,24 @@ All notable changes to MCP Memory are documented here. The format is based on [K
 
 ## [Unreleased]
 
+## [1.23.0] - 2026-10-12
+
+### Added
+
+- The compatibility stdio package now exposes the canonical server's resources and prompts, including shared guidance and complete-result artifact reads, alongside tool calls.
+- Read complete oversized Memory results through authenticated artifact continuation and renewable downloads, with separate content and link expiry. Preserve full content, original recovery receipts and honest storage-failure status without repeating tools.
+
+### Fixed
+
+- Clarify CRM work identifiers, mutation prerequisites and probability bounds. Correct conditional write and replay annotations for pipeline creation, default-vault key issuance, Library ingestion, tag aliases and video status; remove the outdated fixed vault count.
+- Hosted branch builds now install their own workspace packages before dependency resolution, so testing a new version does not require publishing it to npm first.
+
+## [1.22.1] - 2026-09-22
+
+### Changed
+
+- Remove Personal Assistant context-packet tools from the hosted Memory MCP catalog while preserving the underlying source and persisted Memory data for future recovery.
+
 ## [1.22.0] - 2026-09-02
 
 ### Changed
@@ -252,7 +270,9 @@ All notable changes to MCP Memory are documented here. The format is based on [K
 - The hosted service now consumes `mcpscraper-memory-db` 1.5 and `mcpscraper-memory-tools` 1.6, aligning direct OAuth and stdio clients with the current 14-vault Smart RAG contract, including bounded multi-query retrieval, graph expansion, and Jina reranking metadata.
 - Server guidance now documents the Skills vault and the exact hybrid retrieval workflow exposed by `memory-search`.
 
-[Unreleased]: https://github.com/VilovietaSEO/mcp-scraper/compare/memory-v1.22.0...HEAD
+[Unreleased]: https://github.com/VilovietaSEO/mcp-scraper/compare/memory-v1.23.0...HEAD
+[1.23.0]: https://github.com/VilovietaSEO/mcp-scraper/compare/memory-v1.22.1...memory-v1.23.0
+[1.22.1]: https://github.com/VilovietaSEO/mcp-scraper/compare/memory-v1.22.0...memory-v1.22.1
 [1.22.0]: https://github.com/VilovietaSEO/mcp-scraper/compare/memory-v1.21.2...memory-v1.22.0
 [1.21.2]: https://github.com/VilovietaSEO/mcp-scraper/compare/memory-v1.21.1...memory-v1.21.2
 [1.21.1]: https://github.com/VilovietaSEO/mcp-scraper/compare/memory-v1.21.0...memory-v1.21.1

package/README.md

@@ -1,19 +1,21 @@
 # mcp-memory
 
-A hosted, multi-tenant MCP server for structured personal/business memory — notes across 15
+A hosted, multi-tenant MCP server for structured personal/business memory — notes across the current
 purpose-built vaults, semantic search, typeahead, and real structured data tables (typed columns,
 filter, sort) — plus **Scheduled Actions**: freeform-text automations that run on a cadence, using
 your [MCP Scraper](https://mcpscraper.dev) tools to gather live data and mcp-memory to store it.
 
 Customer setup uses one endpoint, `https://mcpscraper.dev/mcp`, and one long-lived credential,
 `MCP_SCRAPER_API_KEY`, for scraping and Memory. This package is a compatibility stdio proxy to that
-unified server; nothing runs or is stored on your machine.
+unified server. It forwards tools, resources and prompts, including the shared guidance and
+complete-result artifact reads. The compatibility bridge runs locally; Memory data remains hosted.
 
 This repository also owns the independently publishable `mcpscraper-memory-db` workspace for
 Postgres/PgVector storage, vault contracts, Jina embeddings, and reranking, and the
-`mcpscraper-memory-tools` workspace for public tool behavior. The hosted backend consumes their
-released package versions. Release those dependencies first, then update and redeploy this service
-before downstream products advertise a newer memory-tool contract.
+`mcpscraper-memory-tools` workspace for public tool behavior. The hosted build packages their
+built workspace code into the deployable function, so deploying a branch includes that branch's
+storage and tool implementations without a separate public package release. Package identities
+and public release versions remain independent.
 
 ## Repository development
 
@@ -31,6 +33,12 @@ npm run verify:packages
 Package identities and release versions remain independent. See [`packages/README.md`](packages/README.md)
 for ownership boundaries and import provenance.
 
+`build:server` expects `build:packages` to have completed. It temporarily materializes local
+workspace archives for Mastra's dependency resolution, restores the original dependency links,
+and includes those archives in the Vercel function. The final check compares every installed
+workspace build file with its branch-local source build. Generated function dependencies and
+their lockfile refer to the included archives, so reinstalling them preserves the branch code.
+
 ## Connect
 
 **OAuth remote connector** (Claude.ai web, Claude Desktop's "Add custom connector") — add

package/dist/bin/mcp-stdio-server.js

@@ -3,10 +3,18 @@
 // bin/mcp-stdio-server.ts
 import { Server } from "@modelcontextprotocol/sdk/server/index.js";
 import { StdioServerTransport } from "@modelcontextprotocol/sdk/server/stdio.js";
-import { ListToolsRequestSchema, CallToolRequestSchema } from "@modelcontextprotocol/sdk/types.js";
+import {
+  ListToolsRequestSchema,
+  CallToolRequestSchema,
+  ListResourcesRequestSchema,
+  ListResourceTemplatesRequestSchema,
+  ReadResourceRequestSchema,
+  ListPromptsRequestSchema,
+  GetPromptRequestSchema
+} from "@modelcontextprotocol/sdk/types.js";
 
 // src/version.ts
-var PACKAGE_VERSION = "1.22.0";
+var PACKAGE_VERSION = "1.23.0";
 
 // bin/mcp-stdio-server.ts
 var wantsHelp = process.argv.includes("--help") || process.argv.includes("-h");
@@ -20,7 +28,7 @@ function renderHelp() {
     "",
     `API key: ${status}`,
     "",
-    "This compatibility package forwards every tool call to the canonical",
+    "This compatibility package forwards tools, resources and prompts to the canonical",
     "MCP Scraper server. Scraping and Memory use the same customer API key.",
     "",
     "Setup:",
@@ -42,6 +50,7 @@ if (!apiKey) {
   process.stderr.write("MCP_SCRAPER_API_KEY is required\n");
   process.exit(1);
 }
+var authenticatedApiKey = apiKey;
 var remoteUrl = (process.env.MCP_SCRAPER_MCP_URL?.trim() || "https://mcpscraper.dev/mcp").replace(/\/$/, "");
 var rpcId = 0;
 function parseRpcBody(text) {
@@ -55,7 +64,7 @@ async function rpc(method, params) {
     headers: {
       "Content-Type": "application/json",
       Accept: "application/json, text/event-stream",
-      "x-api-key": apiKey
+      "x-api-key": authenticatedApiKey
     },
     body: JSON.stringify({ jsonrpc: "2.0", id: ++rpcId, method, params })
   });
@@ -80,20 +89,41 @@ async function main() {
 `);
     process.exit(1);
   }
+  const capabilities = {};
+  if (remoteInfo.capabilities?.tools) capabilities.tools = {};
+  if (remoteInfo.capabilities?.resources) capabilities.resources = {};
+  if (remoteInfo.capabilities?.prompts) capabilities.prompts = {};
   const server = new Server(
     { name: "mcp-memory", version: PACKAGE_VERSION },
-    { capabilities: { tools: {} }, instructions: remoteInfo.instructions }
+    { capabilities, instructions: remoteInfo.instructions }
   );
-  server.setRequestHandler(ListToolsRequestSchema, async () => {
-    const result = await rpc("tools/list");
-    return { tools: result.tools };
-  });
-  server.setRequestHandler(CallToolRequestSchema, async (request) => {
-    return await rpc("tools/call", {
-      name: request.params.name,
-      arguments: request.params.arguments
+  if (capabilities.tools) {
+    server.setRequestHandler(ListToolsRequestSchema, async (request) => {
+      return await rpc("tools/list", request.params);
     });
-  });
+    server.setRequestHandler(CallToolRequestSchema, async (request) => {
+      return await rpc("tools/call", request.params);
+    });
+  }
+  if (capabilities.resources) {
+    server.setRequestHandler(ListResourcesRequestSchema, async (request) => {
+      return await rpc("resources/list", request.params);
+    });
+    server.setRequestHandler(ListResourceTemplatesRequestSchema, async (request) => {
+      return await rpc("resources/templates/list", request.params);
+    });
+    server.setRequestHandler(ReadResourceRequestSchema, async (request) => {
+      return await rpc("resources/read", request.params);
+    });
+  }
+  if (capabilities.prompts) {
+    server.setRequestHandler(ListPromptsRequestSchema, async (request) => {
+      return await rpc("prompts/list", request.params);
+    });
+    server.setRequestHandler(GetPromptRequestSchema, async (request) => {
+      return await rpc("prompts/get", request.params);
+    });
+  }
   const transport = new StdioServerTransport();
   await server.connect(transport);
 }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "mcpscraper-memory",
-  "version": "1.22.0",
+  "version": "1.23.0",
   "description": "mcp-memory — a hosted, multi-tenant, access-controlled per-user Obsidian/Neon memory exposed AS an MCP server, with a scheduled optimization layer. Its own standalone MCP product.",
   "type": "module",
   "bin": {
@@ -49,7 +49,7 @@
     "@mastra/pg": "latest",
     "@modelcontextprotocol/sdk": "^1.29.0",
     "mcpscraper-memory-db": "1.18.0",
-    "mcpscraper-memory-tools": "1.19.0",
+    "mcpscraper-memory-tools": "1.19.1",
     "zod": "^3.25.0"
   },
   "devDependencies": {