@squirrelnotes.app/mcp-embeddings 2.1.1-test.343 → 2.1.1-test.345

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/README.md CHANGED
@@ -80,8 +80,18 @@ installs exactly the dependency versions that were audited and licence-checked
80
80
  for the release.
81
81
 
82
82
  **Size.** The package is about 17 MB to download, most of it the model.
83
- `onnxruntime-node` adds about 300 MB unpacked, because it ships binaries for
84
- every platform. That is why embeddings are a separate package.
83
+ `onnxruntime-node` adds about 301 MB (287 MiB) unpacked, because it ships
84
+ binaries for every platform it supports. That is why embeddings are a separate
85
+ package.
86
+
87
+ **Intel Macs.** onnxruntime-node 1.30.0 ships no binary for Intel Macs (darwin/x64):
88
+ its native bindings cover macOS on Apple silicon, Linux x64 and arm64, and
89
+ Windows x64 and arm64. On an Intel Mac the install should still succeed (the
90
+ runtime's install script has nothing to fetch there), but the runtime cannot
91
+ load, so the server logs `memory: local embeddings were not loaded (runtime);
92
+ memory search is lexical only` once on stderr and `search_memory` stays lexical,
93
+ exactly as with the base package. This is read from the runtime's published
94
+ files and its loader; it has not been tested on an Intel Mac.
85
95
 
86
96
  To go back to lexical search, change `args` back to `@squirrelnotes.app/mcp`.
87
97
  The vector file can be deleted at any time; it is rebuilt when needed.
@@ -123,7 +133,7 @@ stderr, never per call, when:
123
133
  - `SQUIRREL_MEMORY_EMBEDDINGS=0` is set;
124
134
  - a model file's size or SHA-256 differs from `model-manifest.json`, or a
125
135
  model file is missing (the model is refused, never loaded);
126
- - ONNX Runtime cannot load on your platform;
136
+ - ONNX Runtime cannot load on your platform (an Intel Mac, for one: see *Intel Macs* above);
127
137
  - the runtime fails during a search (lexical for the rest of the session).
128
138
 
129
139
  While the model is still loading at start-up (a fraction of a second), search
@@ -1,12 +1,12 @@
1
1
  {
2
2
  "name": "@squirrelnotes.app/mcp-embeddings",
3
- "version": "2.1.1-test.343",
3
+ "version": "2.1.1-test.345",
4
4
  "lockfileVersion": 3,
5
5
  "requires": true,
6
6
  "packages": {
7
7
  "": {
8
8
  "name": "@squirrelnotes.app/mcp-embeddings",
9
- "version": "2.1.1-test.343",
9
+ "version": "2.1.1-test.345",
10
10
  "license": "MIT",
11
11
  "dependencies": {
12
12
  "onnxruntime-node": "1.30.0"
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@squirrelnotes.app/mcp-embeddings",
3
- "version": "2.1.1-test.343",
3
+ "version": "2.1.1-test.345",
4
4
  "description": "Opt-in local semantic memory search for the Squirrel Notes MCP (stdio). The embedding model runs on your machine; no memory text is sent anywhere to be embedded.",
5
5
  "type": "module",
6
6
  "repository": {