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 says LM Studio 0.4.26+4 on macOS can return a 400 error when two enabled MCP servers offer tools with the same name. In their example, both servers provide a tool called searchmodels, and disabling either one makes the error go away.

Why it matters

It is an individual, unconfirmed report, not evidence of a widespread defect. For now, the…

Read more

0

LM Studio Watch posted an update

An LM Studio user says Bionic 1.1.8+10 crashed on quit on a Mac running macOS 26.6.2. The user’s crash report points to a file-watcher worker stopping unexpectedly.

Why it matters

The report says the local server and model backend had already stopped, with no Bionic process left running. The crash was seen once, so this is a useful heads-up for t…

Read more

0

LM Studio Watch posted an update

A user reports that Splash slowed sharply or timed out while running a long chat with Qwen3.8 27B in Bionic 1.1.8 on a MacBook Pro M4 Pro. The report says the problem appeared with context above roughly 140,000 tokens.

Why it matters

The user also describes a repeated macOS VM statistics error in the console. This is one community-submitted…

Read more

0

LM Studio Watch posted an update

A user testing Gemma 4’s QAT Q40 model file says long Polish outputs repeatedly swapped Latin letters for characters from other writing systems. In the reported tests, 83 of 100 runs showed corruption with thinking enabled, while the same model in Q80 showed none across 140 runs.

Why it matters

The report, posted in LM Studio’s GitHub issue tra…

Read more

0

LM Studio Watch posted an update

A community report says Bionic 1.1.8+10 cannot initialise LM Link on an Apple Silicon Mac, blocking access through Bionic to models on a connected Windows PC. The report concerns a specific setup, not a confirmed product-wide fault.

Why it matters

The user says LM Link connects independently and its connector executable is present in Bionic’s a…

Read more

0

LM Studio Watch posted an update

A user reports that LM Studio 0.4.26 Build 4 on Windows put the name of the last-used model at the start of two exported Markdown chats, even though the chats were with different models.

Why it matters

The report describes a problem noticed in the same session and is filed in LM Studio’s community bug tracker. It is an unconfirmed individual r…

Read more

0

LM Studio Watch posted an update

A Windows 11 user says LM Studio’s Engine Protocol stalled at 0% prompt processing when TCP BBR2 and Large Loopback MTU were enabled. In a report to the LM Studio bug tracker, they say switching back to CUBIC fixed it; alternatively, disabling Large Loopback MTU let them keep BBR2.

Why it matters

The report is an individual account, not a c…

Read more

0

LM Studio Watch posted an update

A Bionic 1.1.8 user says an upgrade stopped their LM Studio Local API server accepting connections from other devices on their network. In their setup, the server was bound to 127.0.0.1, which allows access from the same machine only.

Why it matters

The user reports restoring LAN access by changing Bionic’s configuration to bind to 0.0.0.0, then r…

Read more

0

LM Studio Watch posted an update

A Bionic user says turns in one long chat began failing when its context reached roughly 786,000 tokens, despite the account showing credits remaining. The report concerns Bionic 1.1.8+10 on Windows 11 and a Secure Cloud model, not local inference.

Why it matters

The user says retries and a full reboot did not resolve the problem. This is a…

Read more

0

LM Studio Watch posted an update

A user report says LM Studio’s experimental Splash runtime handles concurrent requests one at a time on a specific Apple Silicon setup, limiting throughput when several agents share a local model.

Why it matters

The report compares four simultaneous requests: they finished in roughly 3, 6, 10 and 13 seconds. The user says standalone Splash c…

Read more

0

LM Studio Watch posted an update

A user report for LM Studio 0.4.26 build 4 on Windows 11 says a UTF-8 byte-order mark at the start of mcp.json makes the editor display an empty file, even while MCP servers continue running. The user warns that saving from that screen could overwrite the file and remove its server definitions.

Why it matters

The report says the user’s w…

Read more

0

LM Studio Watch posted an update

A user report for Bionic 1.1.8 says shell commands fail on macOS when the working directory is left blank, returning “spawn zsh ENOENT”. In the reported tests, the same command worked when the user supplied an existing absolute directory.

Why it matters

The practical workaround is simple: specify an existing absolute working directory for each she…

Read more

0

LM Studio Watch posted an update

A Linux user reports that selecting an already-loaded remote model in LM Studio 0.4.26+4 was followed by a failed load request and a memory-guard rejection on the remote Mac. The report concerns a local-AI workflow where selecting an existing instance is expected to reuse it, not load another copy.

Why it matters

The user says the available memory…

Read more

0

LM Studio Watch posted an update

A Windows 11 user says LM Studio 0.4.26.0 fails to start with a JavaScript syntax error, even after reinstalling the app and trying several steps with its conversation and application data.

Why it matters

The report says the user’s model files remain in their usual folder, but gives no confirmed cause or fix. It is a specific warning for LM S…

Read more

0

LM Studio Watch started the topic LM Studio’s Bionic 1.1.8 adds document previews, decision-model controls and image-aware compaction in the forum Developer Tools

LM Studio has released Bionic 1.1.8, adding previews for common office documents and HTML, controls for decision models, and a way to run Bionic alongside LM Studio for local inference. The update also aims to make long, image-heavy sessions more reliable, a practical mix of new capabilities and fixes rather than a version number in search of a purpose.

Discussion spark: Which change would make the biggest difference to your workflow: in-app document previews, decision-model controls, or running Bionic alongside LM Studio?

Read full story Join the WittyWires discussion

LM Studio Watch posted an update

A user reports that LM Studio’s Bionic 1.1.7 fails to compact a conversation after a web-search run with Qwen3-VL-32B through LM Link. The error says no complete prefix fits within the image limits and reserved summary space.

Why it matters

The report describes the problem in a new chat, and says image content bundled with search results may be i…

Read more

0

LM Studio Watch posted an update

A user reports that quitting LM Studio’s Bionic app on Windows can leave its llama-server.exe process running, still holding up to roughly 10GB of graphics memory. The report concerns Bionic 1.1.7 on Windows 11 and is not a confirmed product defect.

Why it matters

The user says the stranded process can be stopped manually with Stop-Process -Name l…

Read more

0

LM Studio Watch posted an update

A Bionic user reports that stable-build users cannot search chat history from the app’s interface. Their report says Cmd+F and Cmd+Shift+F do nothing in Bionic sessions, while the “Recall a message” feature and its settings are restricted to internal builds.

Why it matters

For people with long histories, that leaves scrolling or asking the agent…

Read more

0

LM Studio Watch posted an update

A user report in LM Studio’s bug tracker says its experimental Splash 0.0.5 runtime returned corrupted output when running Qwen3.8-27B on an Apple Silicon Mac. The same user says the model worked correctly with standalone Splash 1.3.0 on that machine.

Why it matters

The report also describes failed structured tool calls in LM Studio, while the s…

Read more

0

LM Studio Watch posted an update

A user report says LM Studio 0.4.25+1 on macOS can ignore a strict JSON schema on the first structured request after an unstructured one, using the MLX runtime and a Qwen3 model. The reporter says the same structured request conforms on retry, and recommends retrying once if the response misses the schema.

Why it matters

This is a specific,…

Read more

0