Repository navigation
v2026.9.3
·
285 commits
to main
since this release
OpenClaw Windows Hub v2026.9.3
Downloads
- Installer (x64):
OpenClawCompanion-Setup-x64.exe- Intel/AMD 64-bit - Installer (ARM64):
OpenClawCompanion-Setup-arm64.exe- Windows on ARM (Surface, etc.) - Portable x64:
OpenClawTray-2026.9.3-win-x64.zip - Portable ARM64:
OpenClawTray-2026.9.3-win-arm64.zip
Features
- 🦞 System tray integration with gateway status
- 🔄 Auto-updates from GitHub Releases
- ✅ Code-signed with Azure Artifact Signing
Requirements
- Windows 10 version 1903 or later
- WebView2 Runtime
- OpenClaw gateway running locally
Quick Start
- Run the installer for your architecture
- Launch from Start Menu or system tray
- Right-click tray icon → Settings to configure
What's Changed
- refactor(tray): move Permissions into shared-state MVVM ownership by @bkudiess in #1075
- refactor(tray): extract Hub and tray presentation models by @bkudiess in #1082
- refactor(tray): extract App-owned surface lifetimes by @bkudiess in #1085
- refactor(tray): extract final App orchestration owners by @bkudiess in #1088
- refactor(gateway): extract protocol core owners by @bkudiess in #1095
- fix(mxc): tasklist no longer hangs past sandbox timeout by @shanselman in #1151
- chore(deps): bump github/gh-aw-actions/setup-cli from 0.84.1 to 0.86.1 by @dependabot[bot] in #1142
- fix: bind SSH credential handoffs to owned listeners by @shanselman in #1126
- refactor(connection): extract domain coordinators by @bkudiess in #1101
- refactor(chat): extract runtime state owners by @bkudiess in #1096
- refactor(chat): extract composer ownership by @bkudiess in #1098
- Bump Microsoft.ML.OnnxRuntime from 1.28.0 to 1.29.0 by @dependabot[bot] in #1170
- Bump Microsoft.WindowsAppSDK from 2.3.1 to 2.4.0 by @dependabot[bot] in #1171
- refactor(setup): split setup steps and own keepalive processes by @bkudiess in #1093
- fix(tray): restore icon after Explorer restart by @shanselman in #1179
- feat(connection): enforce gateway protocol compatibility by @bkudiess in #1127
- feat(chat): render user and agent media in native chat by @calebeden in #1119
- fix(release): sign all first-party payload binaries by @shanselman in #1182
- fix: MCP server accepts malformed IPv6 loopback hosts by @joelagnel in #1180
- refactor(chat): remove the retired FunctionalUI chat surface by @calebeden in #1110
- feat: add managed local AI with llama-server by @joelagnel in #1178
- fix(chat): preserve persisted tool activity details by @TheAngryPit in #1111
- fix(setup): require ownership before WSL cleanup by @shanselman in #1229
- Make telemetry test collectors thread-safe by @shanselman in #1226
- fix: gateway controls preserve Linux PATH in WSL by @shanselman in #1227
- build(deps): bump github/gh-aw-actions/setup-cli from 0.86.1 to 0.87.2 by @dependabot[bot] in #1209
- fix(setup): avoid duplicate failure message by @RomneyDa in #1187
- Stabilize accessibility navigation readiness by @shanselman in #1228
- refactor(local-ai): clarify managed-model probe contract by @joelagnel in #1210
- fix(setup): bind cleanup to live WSL registration by @shanselman in #1230
- fix: prevent .NET 10 COM wrapper memory leak in production by @jacobtomlinson in #1215
- fix: preserve .wslconfig encoding by @joelagnel in #1224
- fix: keepalive process inspection can hang after PowerShell exits by @SebTardif in #1168
- fix: preserve Core updates across restart timeouts by @shanselman in #1231
- Add complete MiniMax TTS support by @karkarl in #1233
- chore: use MXC SDK 0.8.0 by @shanselman in #1234
- fix: Local AI fails to start after first setup by @TheAngryPit in #1208
- fix: render Windows notifications as plain text by @joelagnel in #1223
- fix(setup): block unsafe Local AI archive destinations by @vincentkoc in #1238
- fix(setup): stop WSL inspection flake from dead-ending onboarding by @RomneyDa in #1235
- fix(local-ai): bound default llama generation by @jacobtomlinson in #1241
- chore(local-ai): bump llama.cpp runtime to b10655 by @larroy in #1252
- fix(local-ai): prevent routing to an unverified managed model by @joelagnel in #1212
- test(setup): prove archive link cleanup ordering by @vincentkoc in #1225
- fix(local-ai): model setup unavailable after install by @shanselman in #1254
- fix(chat): restore default reasoning after off by @shanselman in #1255
- fix(setup): recover uninitialized WSL automatically by @karkarl in #1202
- fix(chat): prevent stale attachment metadata writes by @shanselman in #1259
- Fix Start menu activation for running tray app by @shanselman in #1262
- Fix onboarding wizard Back desynchronization by @karkarl in #1205
- fix(setup): gate Local AI on WSL readiness by @karkarl in #1203
- test(connection): stabilize listener output waits by @shanselman in #1264
- fix(chat): render Markdown tables by @shanselman in #1260
- chore: automate safe repository triage by @shanselman in #1261
- fix(local-ai): clarify unavailable Local AI setup states (clean replacement for #1204) by @shanselman in #1263
- test: define schedulable Windows proof pools by @shanselman in #1257
- Support upstream stable correction release versions by @karkarl in #1266
- build(deps): bump github/gh-aw-actions/setup-cli from 0.87.2 to 0.87.9 by @dependabot[bot] in #1272
- Add Ollama inference to the Windows node by @shanselman in #1269
- docs: refresh Windows Crabbox guidance by @RomneyDa in #1283
- Bump Microsoft.Windows.SDK.BuildTools from 10.0.28000.2526 to 10.0.28000.2705 by @dependabot[bot] in #1273
- Fix camera/screen capture consent hang for unattended MCP callers by @shanselman in #1300
- fix(test): stabilize hardware and UI integration tests by @karkarl in #1302
- fix(local-ai): bound GPU admission by dedicated device memory by @joelagnel in #1253
- fix(local-ai): recommend Qwen3.8 with measured context profiles by @joelagnel in #1281
- style(setup-ui): simplify Local AI onboarding copy by @joelagnel in #1290
- fix(release): decouple stable correction validation from upstream releases by @shanselman in #1303
- fix(setup): prevent Local AI routing ambiguity after rerun by @joelagnel in #1282
- docs: add OpenClaw global triage skill by @shanselman in #1308
- ci: reduce superseded PR work by @shanselman in #1310
- ci: add fail-closed docs and skill fast path by @shanselman in #1313
- ci: add code impact fast paths by @shanselman in #1315
- Avoid duplicate advanced CodeQL scans by @shanselman in #1318
- feat: add interactive global triage dashboard by @karkarl in #1319
- improve(ci): reduce core NuGet cache payload by @shanselman in #1320
- ci: publish daily alpha releases by @RomneyDa in #1309
- fix(local-ai): surface llama-server root cause on inference failure by @larroy in #1284
- fix(audio): unify Piper extraction deadline by @shanselman in #1294
- fix(setup): enforce one Tailscale probe deadline by @shanselman in #1296
- fix(chat): keep flattened tool results with the correct tool by @shanselman in #1298
- fix(pt-BR): repair Brazilian Portuguese localization (replaces #1265) by @shanselman in #1295
- fix(chat): stop composer render feedback loop by @shanselman in #1326
- fix(mxc): require safe BaseContainer for Windows node sandboxing by @shanselman in #1327
- Make the winnode reference a discoverable agent skill by @RomneyDa in #1346
- fix(setup): prompt to restart Windows after installing WSL by @shanselman in #1347
- fix(setup): let uninstall recover from absent WSL distro by @shanselman in #1348
- fix(setup): keep WSL diagnostics readable with deterministic UTF-8 by @shanselman in #1350
- Bound WebSocket reconnect attempts by @shanselman in #1352
- Avoid repeated WSL check before installation by @shanselman in #1356
- fix(setup): securely bound CLI installer retries by @shanselman in #1297
- Preserve normal setup command output drains by @shanselman in #1361
- build(deps): bump github/gh-aw-actions/setup-cli from 0.87.9 to 0.88.4 by @dependabot[bot] in #1351
- fix(setup): advance managed Node runtime to 24.19.0 by @RomneyDa in #1359
- fix(build): Piper voice preview fails in local builds with "native library could not be loaded" by @Steven23456 in #1354
- docs: improve LocalAi module documentation and agent coding rules by @larroy in #1329
- build(deps): upgrade Microsoft.UI.Reactor to 0.1.0-preview.14 by @azchohfi in #1376
- improve(setup): install latest stable Gateway from npm by @karkarl in #1312
- fix(setup): complete npm latest Gateway landing by @shanselman in #1381
- fix(setup): verify gateway listener service ownership by @shanselman in #1382
- docs: document safe uninstall and dev cleanup by @shanselman in #1383
- feat(shared): add safe Hugging Face cache primitives by @shanselman in #1385
- fix(chat): make snapshot delivery race-free by @shanselman in #1386
- feat(local-ai): add verified legacy model cache migration by @shanselman in #1388
- fix(local-ai): keep routing safe across endpoint restarts by @shanselman in #1390
- feat(local-ai): safely use the standard Hugging Face hub cache by @shanselman in #1397
- feat(voice): add English (GB) Northern English Male Piper voice by @Steven23456 in #1353
- fix(setup): allow Local AI setup on busy GPUs and current Gateway by @joelagnel in #1399
- fix(update): suppress verified ordinary extended-stable notices by @karkarl in #1392
- fix(local-ai): add non-destructive recovery by @shanselman in #1384
- fix(mxc): bound termination and output drain cleanup by @shanselman in #1387
- fix(chat): prevent bubble text truncation by @karkarl in #1400
New Contributors
- @joelagnel made their first contribution in #1180
- @jacobtomlinson made their first contribution in #1215
- @SebTardif made their first contribution in #1168
- @larroy made their first contribution in #1252
- @Steven23456 made their first contribution in #1354
- @azchohfi made their first contribution in #1376
Full Changelog: v2026.7.1...v2026.9.3