v1.2.3
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Seto Elkahfi committed
Aug 19, 2026 at 08:25 UTC
019cd894b047d05d517a4ec5b199d498c9a352a6
11 files changed
+22
-11
Cargo.lock
+1
-1
@@ -3412,7 +3412,7 @@ checksum = "384b8ab6d37215f3c5301a95a4accb5d64aa607f1fcb26a11b5303878451b4fe"
3412
3413
[[package]]
3414
name = "onde"
3415
-version = "1.2.2"
3415
+version = "1.2.3"
3416
dependencies = [
3417
"anyhow",
3418
"cc",
Cargo.toml
+1
-1
@@ -1,6 +1,6 @@
1
[package]
2
name = "onde"
3
-version = "1.2.2"
3
+version = "1.2.3"
4
edition = "2021"
5
description = "On-device inference engine for Apple silicon."
6
license = "MIT OR Apache-2.0"
sdk/dart/CHANGELOG.md
+6
@@ -1,3 +1,9 @@
1
+## 1.2.3
2
+
3
+- UQFF model support: `OndeInference.uqffModelConfig(...)` builds a config for a pre-quantised UQFF export, and `OndeChatEngine.loadUqffModel(...)` loads it. UQFF weights are quantised ahead of time, so loading skips the memory spike of in-situ quantisation
4
+- Android native libraries are now built with cargo-ndk instead of the hand-rolled CMake path, which fixes NDK version comparison and the ordering between the native-lib build and the Gradle tasks that consume it
5
+- Dependency updates
6
+
7
## 1.2.2
8
9
- Moved onto the published onde-mistralrs 0.9.3 crates
sdk/dart/pubspec.yaml
+1
-1
@@ -2,7 +2,7 @@ name: onde_inference
2
description: >-
3
On-device LLM inference for Flutter & Dart. Run Qwen 2.5 models locally
4
with Metal on iOS and macOS, CPU on Android and desktop. No cloud, no API key.
5
-version: 1.2.2
5
+version: 1.2.3
6
repository: https://github.com/ondeinference/onde
7
homepage: https://ondeinference.com
8
documentation: https://ondeinference.com/sdk/dart
sdk/dart/rust/Cargo.lock
+1
-1
@@ -3604,7 +3604,7 @@ checksum = "384b8ab6d37215f3c5301a95a4accb5d64aa607f1fcb26a11b5303878451b4fe"
3604
3605
[[package]]
3606
name = "onde"
3607
-version = "1.2.2"
3607
+version = "1.2.3"
3608
dependencies = [
3609
"anyhow",
3610
"cc",
sdk/kotlin/gradle.properties
+1
-1
@@ -10,7 +10,7 @@ kotlin.code.style=official
10
# Maven publishing coordinates — keep in sync with Cargo.toml [package] version
11
GROUP=com.ondeinference
12
POM_ARTIFACT_ID=onde-inference
13
-VERSION_NAME=1.2.2
13
+VERSION_NAME=1.2.3
14
15
16
sdk/react-native/CHANGELOG.md
+5
@@ -1,3 +1,8 @@
1
+## 1.2.3
2
+
3
+- UQFF model support: `uqffModelConfig(...)` builds a config for a pre-quantised UQFF export, and `OndeChatEngine.loadUqffModel(...)` loads it. UQFF weights are quantised ahead of time, so loading skips the memory spike of in-situ quantisation
4
+- Dependency updates
5
+
6
## 1.2.2
7
8
- Moved onto the published onde-mistralrs 0.9.3 crates
sdk/react-native/package-lock.json
+2
-2
@@ -1,12 +1,12 @@
1
{
2
"name": "@ondeinference/react-native",
3
- "version": "1.2.2",
3
+ "version": "1.2.3",
4
"lockfileVersion": 3,
5
"requires": true,
6
"packages": {
7
"": {
8
"name": "@ondeinference/react-native",
9
- "version": "1.2.2",
9
+ "version": "1.2.3",
10
"license": "MIT OR Apache-2.0",
11
"devDependencies": {
12
"expo-modules-core": "^2.0.0",
sdk/react-native/package.json
+1
-1
@@ -1,6 +1,6 @@
1
{
2
"name": "@ondeinference/react-native",
3
- "version": "1.2.2",
3
+ "version": "1.2.3",
4
"description": "On-device LLM inference for React Native. Run Qwen 2.5 models locally with Metal on iOS, CPU on Android. No cloud, no API key.",
5
"main": "build/index.js",
6
"types": "build/index.d.ts",
sdk/react-native/rust/Cargo.lock
+2
-2
@@ -3420,7 +3420,7 @@ checksum = "384b8ab6d37215f3c5301a95a4accb5d64aa607f1fcb26a11b5303878451b4fe"
3420
3421
[[package]]
3422
name = "onde"
3423
-version = "1.2.2"
3423
+version = "1.2.3"
3424
dependencies = [
3425
"anyhow",
3426
"cc",
@@ -3777,7 +3777,7 @@ dependencies = [
3777
3778
[[package]]
3779
name = "onde-react-native"
3780
-version = "1.2.2"
3780
+version = "1.2.3"
3781
dependencies = [
3782
"jni",
3783
"log",
sdk/react-native/rust/Cargo.toml
+1
-1
@@ -1,6 +1,6 @@
1
[package]
2
name = "onde-react-native"
3
-version = "1.2.2"
3
+version = "1.2.3"
4
edition = "2021"
5
publish = false
6