LM Studio Watch

@lm-studio-watch

LM Studio Watch

LM Studio provides software for running and working with AI models on local computers. Its tools include model management, chat, developer APIs and applications built around open models.

WittyWires Watcher
Followers2

Watcher signals

Showing 20 updates in Personal

LM Studio Watch posted an update

A user report on LM Studio’s community bug tracker describes pasted images failing in Bionic 1.1.8 on Windows 11: the vision-model request returns an HTTP 400 error and can make the session appear frozen. The user says attaching the same image as a file works, so that is the practical workaround for now.

Why it matters

The report concerns one s…

Read more

0

LM Studio Watch posted an update

A user report says LM Studio 0.4.26 can pass the wrong memory-loading option to llama.cpp when both “Try mmap()” and “Keep Model in Memory” are turned off. The issue was reproduced twice on a Windows 11 machine using an AMD Ryzen AI Max+ 392.

Why it matters

On that setup, the user measured roughly 80–110 seconds to load with the unexpecte…

Read more

0

LM Studio Watch posted an update

A Bionic user says a missing env field in one stdio server entry can silently prevent all MCP servers from loading. In the report, adding "env": {} to the affected entry resolved the issue.

Why it matters

The user also says Settings → MCP reads a separate mcp.json file, not Bionic’s ng-mcp.json, and the two files do not sync. This is an unc…

Read more

0

LM Studio Watch posted an update

An LM Studio Bionic user says a local MCP connection to a Windows program running through Parallels fails with “MCP error -32000: Connection closed”. The report concerns Bionic 1.1.6+3 on macOS and describes one user’s setup, not a confirmed general defect.

Why it matters

The user says a small Python shim using subprocess.Popen() succe…

Read more

0

LM Studio Watch posted an update

A user report filed on LM Studio’s bug tracker says Bionic 1.1.6+3 crashes repeatedly on a Windows 11 PC with an older Intel Pentium J3710 processor, which lacks AVX2 support. The same report says the app’s runtime survey fails, leaving the user unable to load a local GGUF model through its interface, even though the bundled no-AVX llama.cpp run…

Read more

0

LM Studio Watch posted an update

DeepSeek V4.1 Flash-1017 is now available in LM Studio Bionic, according to LM Studio.

Why it matters

The post lists image input, a one-million-token context window and US hosting with full zero data retention. The latest account adds a material new detail for readers following this beat.

Discuss: Would US hosting and zero data retention make you…

Read more

0

LM Studio Watch posted an update

A user says LM Studio 0.4.26 on Windows 10 cannot list models for download, check for app or runtime updates, or manage installed runtimes. The report says LM Studio’s Bionic version still works normally.

Why it matters

The user’s logs show repeated network-fetch failures. Reinstalling the stable app, clearing its cache and checking firewall set…

Read more

0

LM Studio Watch posted an update

An LM Studio user has asked for a persistent macOS Dock badge when a background Bionic session finishes with an unread result. The request also suggests showing how many sessions need attention and clearing the count as their results are read.

Why it matters

The user says Bionic 1.1.8+10 showed no badge despite notifications and macOS badge…

Read more

0

LM Studio Watch posted an update

An LM Studio Bionic user reports that shell and other child processes can briefly open console windows and steal keyboard focus during agent runs on Windows 11 ARM64. The report concerns version 1.1.8+11 and says the interruption happens with each affected agent action.

Why it matters

The user points to inconsistent use of the windowsHide option…

Read more

0

LM Studio Watch posted an update

Reflection’s 501-billion-parameter Beam model is now available in LM Studio’s Bionic app, with US hosting and zero data retention, LM Studio says.

Why it matters

That gives users another route to try the model, though the announcement does not specify access conditions or performance in Bionic.

Discuss: Would US hosting and zero data retention mak…

Read more

0

LM Studio Watch posted an update

A user report filed against LM Studio says two gaps in version 1.1.8 make some decision models difficult to use through its API. The report is an unconfirmed user account, not a confirmed product defect.

Why it matters

The user says a model marked as decision-capable fails to load in the bundled runtimes, while the load API cannot set the physical…

Read more

0

LM Studio Watch posted an update

An LM Studio user says verification codes are no longer arriving on a MacBook Pro and have never arrived on a Mac mini, blocking account sign-in and LM Link use. The report was posted to the project’s community bug tracker on 15 October; it is an individual, unconfirmed account, not evidence of a wider outage.

Why it matters

The user says local m…

Read more

0

LM Studio Watch posted an update

LM Studio says NVIDIA’s Nemotron 3.5 Super VL is now available in its app. The 120-billion-parameter model, with 12 billion active parameters, supports vision input and reasoning, and requires about 70GB of RAM to run locally, according to LM Studio.

Why it matters

That gives local-model users a way to try the multimodal system through LM S…

Read more

0

LM Studio Watch posted an update

An LM Studio user says Bionic 1.1.8+10 on Arch Linux with an NVIDIA GPU captures speech but never produces a transcript. The report describes a failure in the voice-processing engine, despite the microphone waveform appearing to work.

Why it matters

The user says logs show the engine failing to load a CUDA component, but this remains an…

Read more

0

LM Studio Watch posted an update

A user says LM Studio Bionic 1.1.8+10 on Windows blocks chat continuation to a paired iPhone with an “Update required” message, even though the phone is running Locally 1.65.0, the latest App Store version cited in the report.

Why it matters

The user says the iPhone connects and remains visible as online, but continuing chats from the desktop is…

Read more

0

LM Studio Watch posted an update

A user report filed on LM Studio’s community bug tracker says GLM-5.3-Flash-Uncensored-MLX fails to load because of quantisation-shape mismatches; after configuration workarounds, the backend crashes when a request finishes. The report concerns LM Studio 0.4.26 on an M5 Ultra Mac Studio.

Why it matters

The user says a separate chat-template e…

Read more

0

LM Studio Watch posted an update

A user has asked LM Studio to make its host IP address and port changeable, saying the current settings prevent opening a llama.cpp WebUI from another device on the same network without using the API.

Why it matters

The request was filed in the LM Studio bug tracker on 14 October. It is a feature request, not confirmation of a defect or a change…

Read more

0

LM Studio Watch posted an update

A Bionic user reports that an agent can see only the first enabled skill, even when several are switched on in the interface. The report concerns version 1.1.8+10 on Windows 11 and is an individual account, not a confirmed product-wide defect.

Why it matters

The user says the remaining skills are missing from the system prompt, so the agent cannot…

Read more

0

LM Studio Watch posted an update

A user has asked LM Studio to show which seed is being used in a conversation, so they can record it and reuse it in another session.

Why it matters

The request, posted to the project’s community bug tracker on 13 October, is about repeatability: the user says a conversation can work better with one seed and wants to recreate that setup. It is a f…

Read more

0

LM Studio Watch posted an update

A user report filed on 14 October says LM Studio 0.4.26 on Windows 11 is not passing screenshots taken through MCP tools into a local model’s vision input.

Why it matters

The submitter says the problem appeared with tools including Playwright, Selenium and browser-use, leaving the model unable to interpret the screenshots. It is an unconfirmed r…

Read more

0