v0.1.801-beta

Thanks for the support for Qwen3.8-27B and Unsloth Desktop last week! For this release, we merged 200+ PRs to introduce many new features, fixes including:

* **Auto Compaction (Experimental)** for longer chats beyond context limits
* **Remote & LAN Access (Preview)** for easy network access without Cloudflare links
* **Faster Chat** - Improved streaming performance, reduced UI lag, and smoother long conversations.
* Support for **custom llama.cpp builds**. Toggles for Cache RAM, Mmap, Mlock, Checkpoints, Spe Decoding KV Cache, Vision On / Off
* [Unsloth Dynamic v3.0](https://unsloth.ai/docs/basics/dynamic-3.0-ggufs) is released. New Qwen3.8-27B Dynamic v3.0 GGUFs deliver >10% higher top-1 accuracy compared to everyone else. Works with Unsloth.

## Auto compaction (Experimental)
You can now do long chats and surpass a model's context limit, with evicted turns
remaining searchable.
* Long local chats can continue past the context limit by rolling older turns out of
  active context.
* Auto compaction creates fresh context epochs instead of permanently trimming replies.
* Evicted conversations remain searchable, with improved fact recall.
* Context window information is shown before chat begins.

**How it works**
* Entire oldest turns are removed only when needed, never mid-message. The saved transcript stays unchanged.
* Evicted turns are indexed into a per-thread searchable archive using Unsloth's existing RAG pipeline (store, chunker, embedder, retrieval). Lexical search is prioritized because chat recall is usually exact matches like names, numbers, or IDs.
* A recall is forced during eviction instead of relying on the model to search itself. Later retrieval uses `search_conversation`.
* The archive persists across epochs, allowing future compactions to recover previously evicted context.
* No summarization: it showed little benefit and added ~190s per compaction.

## Remote & LAN access (Preview)
Access Unsloth from another device on your network, managed from settings.
* New dedicated Settings section for remote access.
* Enable or disable LAN access without restarting.
* Supports connection addresses, QR codes, and optional auto-start.
* LAN access is disabled by default and requires changing the generated admin password.

## Chat + Functionality improvements

* Long chats are faster with optimized streaming and thread handling.
* Projects organize chats, files, and workspaces together.
* Chats remember composer settings, system prompts, and model sampling options.
* Added prompt queueing and customizable keyboard shortcuts.
* Added `edit_file` tool for partial file modifications.
* Improved tool calling and MCP behavior.

## Hardware + inference improvements

* Support for custom llama.cpp builds.
* Toggles for advanced settings - Cache RAM, Mmap, Mlock, Checkpoints, Speculative Decoding KV Cache, Vision On / Off
* Better validation for missing GPU support and oversized GGUF models.
* Improved MTP performance and VRAM handling.
* Better compatibility across ROCm, xFormers, and flash-attention setups.
* Added Intel XPU support.

## API + serving

* Responses API supports structured output formatting.
* Improved recovery when llama-server crashes.
* Added OpenCode V2 support for `unsloth start`.

## Hub + Desktop + training

* Hub supports more image/video workflows.
* Faster and improved GGUF downloads.
* Desktop packaging and permissions improved.
* Training scheduler and preprocessing improvements.

<table>
  <tr>
    <td><b>Platform</b></td>
    <td><b>Link</b></td>
  </tr>
  <tr>
    <td><b>Windows</b></td>
    <td><a href='https://github.com/unslothai/unsloth/releases/download/v0.1.801-beta/Unsloth-Desktop-0_1_801_beta-Windows.exe'>Download</a></td>
  </tr>
  <tr>
    <td><b>macOS</b></td>
    <td><a href='https://github.com/unslothai/unsloth/releases/download/v0.1.801-beta/Unsloth-Desktop-0_1_801_beta-MacOS.dmg'>Download</a></td>
  </tr>
  <tr>
    <td><b>Linux / Ubuntu (deb)</b></td>
    <td><a href='https://github.com/unslothai/unsloth/releases/download/v0.1.801-beta/Unsloth-Desktop-0_1_801_beta-Ubuntu.deb'>Download</a></td>
  </tr>
  <tr>
    <td><b>Linux (AppImage)</b></td>
    <td><a href='https://github.com/unslothai/unsloth/releases/download/v0.1.801-beta/Unsloth-Desktop-0_1_801_beta-Linux.AppImage'>Download</a></td>
  </tr>
  <tr>
    <td><b>Linux (Arm64)</b></td>
    <td><a href='https://github.com/unslothai/unsloth/releases/download/v0.1.801-beta/Unsloth-Desktop-0_1_801_beta-ARM64.app.tar.gz'>Download</a></td>
  </tr>
</table>

## What's Changed
* Bump install.sh / install.ps1 pin to unsloth>=2026.8.18 by @danielhanchen in https://github.com/unslothai/unsloth/pull/8836
* Give the llama-flag fixtures the parallelSlotsClamped they now require by @vineethsaivs in https://github.com/unslothai/unsloth/pull/8837
* Studio: preserve pre-tool reasoning inside the GGUF tool loop by @losuler in https://github.com/unslothai/unsloth/pull/8581
* docs: fix dead Linux .deb download link by @yzxcj797 in https://github.com/unslothai/unsloth/pull/8891
* Model hub: show the Meta mark on the Unsloth re-upload of Muse Glimmer by @shimmyshimmer in https://github.com/unslothai/unsloth/pull/8949
* Studio: speed up the local model inventory and keep it off the API loop by @oobabooga in https://github.com/unslothai/unsloth/pull/8770
* Studio: reuse the cached GGUF work a load request has already done by @oobabooga in https://github.com/unslothai/unsloth/pull/8771
* studio: keep each tool-call round separate when the delta index restarts by @mahiatlinux in https://github.com/unslothai/unsloth/pull/8754
* studio: route bare tool-call fragments to the call that owns the index now by @mahiatlinux in https://github.com/unslothai/unsloth/pull/8755
* studio: opt-in model auto-switch for the image and video apis by @mahiatlinux in https://github.com/unslothai/unsloth/pull/8766
* Studio: keep project chats in Recents, add a Projects sidebar section by @shimmyshimmer in https://github.com/unslothai/unsloth/pull/8932
* Studio: find the real Documents folder on Windows for project workspaces by @shimmyshimmer in https://github.com/unslothai/unsloth/pull/8955
* fix(unsloth studio): only preprocess the rows a max_steps run will actually use instead of preprocessing validated datasets by @LeoBorcherding in https://github.com/unslothai/unsloth/pull/8890
* Studio: stop the chat UI falling behind a fast stream by @oobabooga in https://github.com/unslothai/unsloth/pull/8845
* studio: keep each chat's composer pills and settings with the chat by @mahiatlinux in https://github.com/unslothai/unsloth/pull/8686
* studio: allow a per-connection max tokens limit on every provider by @mahiatlinux in https://github.com/unslothai/unsloth/pull/8944
* Crash guard: read dumpability in order, and stop losing crashes to scope and naming by @danielhanchen in https://github.com/unslothai/unsloth/pull/8808
* Studio: size the max_steps row bound for the data-parallel world by @danielhanchen in https://github.com/unslothai/unsloth/pull/8958
* Studio: make the paste-to-attachment threshold a chat setting by @shimmyshimmer in https://github.com/unslothai/unsloth/pull/8963
* Fix CI on main: stale test doubles, a stale router stub, and two source defects by @danielhanchen in https://github.com/unslothai/unsloth/pull/8956
* Make the startup profile a gate, with budgets from its own measurements by @danielhanchen in https://github.com/unslothai/unsloth/pull/8965
* Keep xFormers working when flash-attn 4 is installed, and guard the varlen int32 overflow by @danielhanchen in https://github.com/unslothai/unsloth/pull/8957
* Stop six backend tests waiting on the clock instead of on a signal by @danielhanchen in https://github.com/unslothai/unsloth/pull/8975
* CI: stop three workflows paying for work they throw away by @danielhanchen in https://github.com/unslothai/unsloth/pull/8976
* Give the pasted-text import an extension so the node test runner can load it by @danielhanchen in https://github.com/unslothai/unsloth/pull/8981
* Make the packed-boundary guard reachable on the fused cross-entropy path by @danielhanchen in https://github.com/unslothai/unsloth/pull/8959
* Studio: Fix embedded MTP performance under partial GPU offload by @oobabooga in https://github.com/unslothai/unsloth/pull/8875
* Studio: tokenize the dataset online for plain-text single-pass runs by @danielhanchen in https://github.com/unslothai/unsloth/pull/8960
* Studio: keep K and V equal when the flash-attention retry resets the V cache by @danielhanchen in https://github.com/unslothai/unsloth/pull/8827
* Studio: hold the Auto-layers context exemption only while --fit runs by @danielhanchen in https://github.com/unslothai/unsloth/pull/8825
* Studio: stop a sent prompt staying in the chat composer by @shimmyshimmer in https://github.com/unslothai/unsloth/pull/8849
* Chat: queue prompts with Cmd/Ctrl+Enter and drag to reorder the queue by @shimmyshimmer in https://github.com/unslothai/unsloth/pull/8952
* Studio: repin the permission-mode contract to behaviour, not a declaration by @danielhanchen in https://github.com/unslothai/unsloth/pull/8986
* Cut Repo tests (CPU) roughly in half: stop the synthetic vLLM tests reaping a fake server by @danielhanchen in https://github.com/unslothai/unsloth/pull/8993
* Core: sweep every transformers model_type across a process pool by @danielhanchen in https://github.com/unslothai/unsloth/pull/8995
* Repair the Tauri retry test against the download-marker pipeline by @danielhanchen in https://github.com/unslothai/unsloth/pull/8996
* Studio: read every launcher world size when resolving a step-capped run's passes by @danielhanchen in https://github.com/unslothai/unsloth/pull/9000
* Repair the prompt-queue contract test against the queueing refactor by @danielhanchen in https://github.com/unslothai/unsloth/pull/9026
* studio: stop prompting to stop chats for a model already loaded by @mahiatlinux in https://github.com/unslothai/unsloth/pull/8943
* Studio: remember chat parameters per model by @shimmyshimmer in https://github.com/unslothai/unsloth/pull/8757
* Studio: share attached files across a project by @shimmyshimmer in https://github.com/unslothai/unsloth/pull/8756
* Fix a 500 on dictation Unload before any backend is resident by @danielhanchen in https://github.com/unslothai/unsloth/pull/9013
* Run the repo CPU test suite on all four runner cores by @danielhanchen in https://github.com/unslothai/unsloth/pull/9019
* Scan package archives across cores instead of one at a time by @danielhanchen in https://github.com/unslothai/unsloth/pull/9024
* Stop TestProbeGating asserting a tier the mapping check is meant to upgrade by @danielhanchen in https://github.com/unslothai/unsloth/pull/9029
* Stub the implicit STT load in the disconnect-cancellation test by @danielhanchen in https://github.com/unslothai/unsloth/pull/9031
* Stop test_begin_load_never_refuses_auto racing the thread it is asserting about by @danielhanchen in https://github.com/unslothai/unsloth/pull/9034
* Measure the streaming stripper against CPU time, not wall clock by @danielhanchen in https://github.com/unslothai/unsloth/pull/9035
* Fix silent lr_warmup_steps no-op under the default constant scheduler by @Eoin-Houstoun in https://github.com/unslothai/unsloth/pull/8464
* Repair the contract tests #8943 left behind on main by @danielhanchen in https://github.com/unslothai/unsloth/pull/9067
* Take the loaded-models indicator suite off the UI job's critical path by @danielhanchen in https://github.com/unslothai/unsloth/pull/9060
* Keep the killed-by-signal exit status when release() fails by @danielhanchen in https://github.com/unslothai/unsloth/pull/9072
* Put the event-loop latency bound between the two things it separates by @danielhanchen in https://github.com/unslothai/unsloth/pull/9077
* Stub unsloth in test_safetensors_reasoning_stream so it can run first by @danielhanchen in https://github.com/unslothai/unsloth/pull/9027
* Report what the launcher logged when a signal test fails by @danielhanchen in https://github.com/unslothai/unsloth/pull/9079
* Studio: let Allow microphone recover from a saved "Don't allow" on Windows by @shimmyshimmer in https://github.com/unslothai/unsloth/pull/9006
* Run both ends of the interpreter matrix on a pull request, all four on main by @danielhanchen in https://github.com/unslothai/unsloth/pull/9080
* fix(studio): enable microphone/camera access in Linux WebKitGTK webview by @rsd-darshan in https://github.com/unslothai/unsloth/pull/8720
* Studio: budget the JavaScript that runs before the first screen by @danielhanchen in https://github.com/unslothai/unsloth/pull/8964
* fix(studio): derive disk percentage from displayed usage by @dyuan311 in https://github.com/unslothai/unsloth/pull/9021
* Studio: stop rereading the whole reply on every streamed arrival by @danielhanchen in https://github.com/unslothai/unsloth/pull/9012
* Studio: stop every message part re-rendering on each streaming chunk by @danielhanchen in https://github.com/unslothai/unsloth/pull/9014
* Studio: stop a LaTeX rewrite throwing away the whole retained markdown prefix by @danielhanchen in https://github.com/unslothai/unsloth/pull/9017
* Studio: stop later code spans changing whether an earlier one is escaped by @danielhanchen in https://github.com/unslothai/unsloth/pull/9028
* Chat: stop a message delete from re-rendering the whole thread by @danielhanchen in https://github.com/unslothai/unsloth/pull/9042
* Chat: stop a keystroke in the composer costing one pass over the thread by @danielhanchen in https://github.com/unslothai/unsloth/pull/9054
* Let every main run finish instead of cancelling it before it starts by @danielhanchen in https://github.com/unslothai/unsloth/pull/9082
* One multi-turn smoke script for all three operating systems by @danielhanchen in https://github.com/unslothai/unsloth/pull/9086
* Studio: give a CRLF reply the same incremental rendering as an LF one by @danielhanchen in https://github.com/unslothai/unsloth/pull/9088
* Studio: sweep the streaming markdown corpus at every prefix by @danielhanchen in https://github.com/unslothai/unsloth/pull/9085
* Do not let the signal handler's own logging stop it dying by @danielhanchen in https://github.com/unslothai/unsloth/pull/9083
* Stop the media switch test racing a 0.3s budget against 2s of cold path by @danielhanchen in https://github.com/unslothai/unsloth/pull/9097
* Studio: honor the Responses API text.format structured output by @NilayYadav in https://github.com/unslothai/unsloth/pull/8846
* Studio: stop the streaming prefix checks scanning the whole reply by @danielhanchen in https://github.com/unslothai/unsloth/pull/9038
* Studio: stop the streamed reply being flattened on every arrival by @danielhanchen in https://github.com/unslothai/unsloth/pull/9049
* Studio: strip the trailing template placeholder once, on the finished reply by @danielhanchen in https://github.com/unslothai/unsloth/pull/9101
* Studio: fix the Windows-only path failures in the frontend test suite, and catch the shape from Linux by @danielhanchen in https://github.com/unslothai/unsloth/pull/8980
* Studio: run the frontend unit tests on Windows in CI by @danielhanchen in https://github.com/unslothai/unsloth/pull/9099
* Run the Backend CI matrix in parallel, like its sibling job already does by @danielhanchen in https://github.com/unslothai/unsloth/pull/9095
* Pair the diffusion warmup presets with a scheduler that uses them by @vineethsaivs in https://github.com/unslothai/unsloth/pull/8593
* Make the models a defaults file claims actually load it by @vineethsaivs in https://github.com/unslothai/unsloth/pull/8726
* Derive warmup steps from a model default that gives a ratio by @vineethsaivs in https://github.com/unslothai/unsloth/pull/8723
* Studio: let the hub run non-GGUF image and video models by @NilayYadav in https://github.com/unslothai/unsloth/pull/8855
* Restore crypto.randomUUID before the Studio bundle evaluates by @Lyxot in https://github.com/unslothai/unsloth/pull/9075
* studio: refuse a gguf that cannot fit in free vram plus available ram by @mahiatlinux in https://github.com/unslothai/unsloth/pull/8883
* studio: add settings-managed LAN access by @mahiatlinux in https://github.com/unslothai/unsloth/pull/8951
* Ask the repo, not sys.path, whether a stubbed name is third-party by @danielhanchen in https://github.com/unslothai/unsloth/pull/9133
* Stop the sidecar tests writing into the real Studio install by @danielhanchen in https://github.com/unslothai/unsloth/pull/9131
* Delete a test that never ran, and stop another appearing by @danielhanchen in https://github.com/unslothai/unsloth/pull/9135
* Make a hung kaggle launcher say where it is stuck by @danielhanchen in https://github.com/unslothai/unsloth/pull/9104
* Stop the banner layout suite waiting out the update timer on every boot by @danielhanchen in https://github.com/unslothai/unsloth/pull/9105
* Give each xdist worker its own torch.compile cache directory by @danielhanchen in https://github.com/unslothai/unsloth/pull/9134
* One interpreter leg on a pull request, and a floor lint that reads more than syntax by @danielhanchen in https://github.com/unslothai/unsloth/pull/9100
* Split Chat UI Tests into four shards on the Studio boundaries by @danielhanchen in https://github.com/unslothai/unsloth/pull/9132
* Bundle Unsloth GGUF CI onto one runner, matching Windows and macOS by @danielhanchen in https://github.com/unslothai/unsloth/pull/9139
* Give the small Windows checks one box per image instead of one box per check by @danielhanchen in https://github.com/unslothai/unsloth/pull/9143
* Let tests skip the settle wait between retried VRAM reads by @danielhanchen in https://github.com/unslothai/unsloth/pull/9141
* Studio: keep the streaming render harness the perf PRs kept rebuilding by @danielhanchen in https://github.com/unslothai/unsloth/pull/8969
* Studio: pin why a Streamdown remount keeps its highlighted code by @danielhanchen in https://github.com/unslothai/unsloth/pull/9048
* Chat: cover which messages a delete takes with it by @danielhanchen in https://github.com/unslothai/unsloth/pull/9087
* Studio: pin what the trailing-placeholder window does with nested placeholders by @danielhanchen in https://github.com/unslothai/unsloth/pull/9091
* Studio: stop a settings module reading a chat store key before it exists by @danielhanchen in https://github.com/unslothai/unsloth/pull/8979
* Studio: load the settings tab panels when they are shown, not at launch by @danielhanchen in https://github.com/unslothai/unsloth/pull/8966
* Name the encoding when reading wslinfo's output by @danielhanchen in https://github.com/unslothai/unsloth/pull/9142
* Studio: keep pandas out of the backend startup import graph by @danielhanchen in https://github.com/unslothai/unsloth/pull/8962
* Run the three loaded-models-indicator engines at once by @danielhanchen in https://github.com/unslothai/unsloth/pull/9158
* Unblock Mac Studio GGUF CI, red on every main run since #8883 by @danielhanchen in https://github.com/unslothai/unsloth/pull/9155
* Drop the Exa preset from the MCP menu by @shimmyshimmer in https://github.com/unslothai/unsloth/pull/9160
* Studio: stop the overlay rail going click-through while it is scrolling by @danielhanchen in https://github.com/unslothai/unsloth/pull/9150
* Open the connection form when Settings > Connections has nothing to list by @shimmyshimmer in https://github.com/unslothai/unsloth/pull/9157
* Security audit: pin openai, and re-review the four digest-pinned entries by @danielhanchen in https://github.com/unslothai/unsloth/pull/9148
* Say why a Xet download sits at 0% until it finishes by @shimmyshimmer in https://github.com/unslothai/unsloth/pull/9159
* Stop running every macOS workflow on every commit to main by @danielhanchen in https://github.com/unslothai/unsloth/pull/9174
* Time the install from CI, without changing the installers by @danielhanchen in https://github.com/unslothai/unsloth/pull/9153
* Studio: measure where a heavy thread stalls, across engines and thread size by @danielhanchen in https://github.com/unslothai/unsloth/pull/9016
* Studio CI: make a failing browser smoke say why, and stop it skipping the build gates by @danielhanchen in https://github.com/unslothai/unsloth/pull/8983
* Stop spending the Actions cache budget on caches nothing reads by @danielhanchen in https://github.com/unslothai/unsloth/pull/9151
* Fix low VRAM mmproj fallback in Studio by @wasimysaid in https://github.com/unslothai/unsloth/pull/9173
* Run two short jobs as background lanes of Lint CI, not on their own runners by @danielhanchen in https://github.com/unslothai/unsloth/pull/9176
* Studio: fail fast when the installed llama.cpp prebuilt has no kernels for this GPU by @NilayYadav in https://github.com/unslothai/unsloth/pull/8841
* Studio: bulk chat management multi-select move/pin/archive/export/delete by @NilayYadav in https://github.com/unslothai/unsloth/pull/8745
* hide app-only MCP tools from the model and explain why local commands are disabled by @NilayYadav in https://github.com/unslothai/unsloth/pull/8739
* Studio: remember last used model in the backend, not per-browser by @NilayYadav in https://github.com/unslothai/unsloth/pull/8587
* Studio: show MCP server display names instead of internal ids by @NilayYadav in https://github.com/unslothai/unsloth/pull/8557
* Studio: make the chat thread stop getting slower as it fills by @danielhanchen in https://github.com/unslothai/unsloth/pull/8992
* Studio: recover marked reports across reasoning channels by @alkinun in https://github.com/unslothai/unsloth/pull/9119
* show the local model's context window before the first token count by @mahiatlinux in https://github.com/unslothai/unsloth/pull/8882
* studio: stop nudging a turn that asks the user for a missing detail by @mahiatlinux in https://github.com/unslothai/unsloth/pull/8946
* Say both load fallbacks when both fire, and repair the two suites that guard it by @danielhanchen in https://github.com/unslothai/unsloth/pull/9189
* Run the Mac GGUF inference phases in the Mac UI job, not on a second runner by @danielhanchen in https://github.com/unslothai/unsloth/pull/9213
* Print what the server said when an inference smoke request 4xx's by @danielhanchen in https://github.com/unslothai/unsloth/pull/9202
* Allowlist huggingface-hub's http_backoff loop so Security audit stops failing by @danielhanchen in https://github.com/unslothai/unsloth/pull/9252
* Count a capped overlay stack by what it reads, not by how the cap is spelled by @danielhanchen in https://github.com/unslothai/unsloth/pull/9253
* Studio: support llama.cpp semantic version output by @oobabooga in https://github.com/unslothai/unsloth/pull/9127
* Bound every apt step in CI, and fix the retry that never ran by @danielhanchen in https://github.com/unslothai/unsloth/pull/9256
* Make apt fail fast, so the retry has something to retry by @danielhanchen in https://github.com/unslothai/unsloth/pull/9260
* Record the venv own requirement digests, not the installers by @danielhanchen in https://github.com/unslothai/unsloth/pull/9263
* Stub all three fields the snapshot reads, not two of them by @danielhanchen in https://github.com/unslothai/unsloth/pull/9274
* Ask whether Playwright system libraries are missing before installing them by @danielhanchen in https://github.com/unslothai/unsloth/pull/9273
* Stop blaming the recipe for a turn that never came back by @danielhanchen in https://github.com/unslothai/unsloth/pull/9266
* Move the two version-incidental macOS jobs onto the emptier pool by @danielhanchen in https://github.com/unslothai/unsloth/pull/9277
* Give the Colab freeze a pip cache, like the job beside it already has by @danielhanchen in https://github.com/unslothai/unsloth/pull/9276
* Wait for the Compare nav instead of sampling for it by @danielhanchen in https://github.com/unslothai/unsloth/pull/9264
* Install the browser engines each Chat UI shard actually drives by @danielhanchen in https://github.com/unslothai/unsloth/pull/9283
* Update README to include Intel XPU support description for unsloth studio by @JoshuaL3000 in https://github.com/unslothai/unsloth/pull/9250
* Allow unlimited Deep Research time budgets by @alkinun in https://github.com/unslothai/unsloth/pull/9078
* Stop the orphan-scan tests asking the host whether an invented PID is alive by @danielhanchen in https://github.com/unslothai/unsloth/pull/9280
* Repair the tests main is currently red on by @danielhanchen in https://github.com/unslothai/unsloth/pull/9192
* Re-type the native default, not whatever the input happens to show by @danielhanchen in https://github.com/unslothai/unsloth/pull/9282
* Studio: confirm an Audio model swap instead of dead-ending on the backend 409 by @shimmyshimmer in https://github.com/unslothai/unsloth/pull/9092
* fix(studio): refuse a hand-set context unified memory cannot hold by @danielhanchen in https://github.com/unslothai/unsloth/pull/9172
* Studio: size Xet download buffers from free RAM, not total by @shimmyshimmer in https://github.com/unslothai/unsloth/pull/9052
* Studio: make the partial download affordance tell the truth by @shimmyshimmer in https://github.com/unslothai/unsloth/pull/8982
* Cache uv's downloads, which is now the largest single cost in CI by @danielhanchen in https://github.com/unslothai/unsloth/pull/9271
* Studio: incrementally tokenize streaming code fences by @oobabooga in https://github.com/unslothai/unsloth/pull/8935
* Desktop: ship a complete Linux AppImage by @oobabooga in https://github.com/unslothai/unsloth/pull/9113
* Studio: link the Read aloud TTS row to Audio and show the preview generate wait by @shimmyshimmer in https://github.com/unslothai/unsloth/pull/9093
* studio: show the model a chat was started on, and offer it back by @shimmyshimmer in https://github.com/unslothai/unsloth/pull/9081
* Desktop: make every drop zone take a drop again (#9036) by @shimmyshimmer in https://github.com/unslothai/unsloth/pull/9056
* Studio: say when a scan folder cannot be read instead of showing no models by @shimmyshimmer in https://github.com/unslothai/unsloth/pull/9053
* Studio: recover the OpenAI passthrough from a crashed llama-server by @shimmyshimmer in https://github.com/unslothai/unsloth/pull/8999
* Studio: stop the desktop health watchdog killing a backend that is busy generating by @shimmyshimmer in https://github.com/unslothai/unsloth/pull/8990
* Studio: add a keyboard shortcuts page and make the existing chords rebindable by @shimmyshimmer in https://github.com/unslothai/unsloth/pull/8948
* Scroll into view before a forced click by @danielhanchen in https://github.com/unslothai/unsloth/pull/9291
* studio: keep each chat's sampling params and system prompt with the chat by @shimmyshimmer in https://github.com/unslothai/unsloth/pull/9055
* Run the zoo suite in parallel, minus the two files that cannot share a worker by @danielhanchen in https://github.com/unslothai/unsloth/pull/9285
* Cache the .deb set webkit needs, so a dead mirror costs a miss not a shard by @danielhanchen in https://github.com/unslothai/unsloth/pull/9289
* Revert "Say why a Xet download sits at 0% until it finishes" by @shimmyshimmer in https://github.com/unslothai/unsloth/pull/9293
* Default preserve thinking on for Qwen3.8 by @Imagineer99 in https://github.com/unslothai/unsloth/pull/9096
* Studio: price a partial GGUF by what is left to fetch by @shimmyshimmer in https://github.com/unslothai/unsloth/pull/8989
* Studio: add an edit_file tool so agents stop rewriting whole files by @shimmyshimmer in https://github.com/unslothai/unsloth/pull/8753
* fix(studio): img2img/inpaint on GGUF image models — from_pipe fallback with no recast (#9186) by @yzxcj797 in https://github.com/unslothai/unsloth/pull/9193
* Install Studio on a Windows host whose C# compiler cannot run by @danielhanchen in https://github.com/unslothai/unsloth/pull/9178
* fix(AMD): retry ROCm llama-server with bundled HIP on HIP/ROCR mismatch by @indrajeetapache in https://github.com/unslothai/unsloth/pull/9002
* Studio:  use a custom llama.cpp build by @Etherll in https://github.com/unslothai/unsloth/pull/9292
* Studio: keep health checks responsive during Anthropic generation by @Imagineer99 in https://github.com/unslothai/unsloth/pull/9201
* Studio: propagate required backend version to repair pipeline (#8610) by @karan68 in https://github.com/unslothai/unsloth/pull/8670
* Keep a sampler's std and uuid format across a save and reopen by @vineethsaivs in https://github.com/unslothai/unsloth/pull/9225
* Studio: never pick a macOS AppleDouble sidecar as a GGUF by @sts-change in https://github.com/unslothai/unsloth/pull/8919
* Studio: follow up #9201 across non-streaming sibling paths by @oobabooga in https://github.com/unslothai/unsloth/pull/9312
* Studio: take common SQLite reads off the event loop thread by @oobabooga in https://github.com/unslothai/unsloth/pull/9234
* Keep xFormers attention masks on the GPU running each layer by @oobabooga in https://github.com/unslothai/unsloth/pull/8516
* Studio: add rolling context windows for local GGUF chat by @alkinun in https://github.com/unslothai/unsloth/pull/8961
* Studio: keep and search the turns rolling context evicts by @danielhanchen in https://github.com/unslothai/unsloth/pull/9074
* Studio: recall the latest version of a fact, not the most quotable one by @danielhanchen in https://github.com/unslothai/unsloth/pull/9161
* Studio: compact a chat by resetting the epoch, not by trimming it forever by @danielhanchen in https://github.com/unslothai/unsloth/pull/9162
* Support OpenCode V2 in unsloth start by @pascalandr in https://github.com/unslothai/unsloth/pull/9275
* Keep path_utils off PEP 604, which the 3.9 floor gate rejects by @danielhanchen in https://github.com/unslothai/unsloth/pull/9335
* Stub the two helpers the sliced harnesses started importing by @danielhanchen in https://github.com/unslothai/unsloth/pull/9342
* Put back the AppleDouble GGUF filters #9074 reverted, and guard them by @danielhanchen in https://github.com/unslothai/unsloth/pull/9353
* Fix the two backend guards that #9292 turned red on the 3.13 leg by @danielhanchen in https://github.com/unslothai/unsloth/pull/9348
* Drain the sampling debounce long enough for the node CI actually runs by @danielhanchen in https://github.com/unslothai/unsloth/pull/9332
* Take two ubuntu jobs off their own runners: absorb one, delete the other by @danielhanchen in https://github.com/unslothai/unsloth/pull/9360
* Studio: stop docked composer from squeezing download panel by @oobabooga in https://github.com/unslothai/unsloth/pull/9334
* Drain the sampling suites on the store's write chains, not on node's loader by @danielhanchen in https://github.com/unslothai/unsloth/pull/9367
* Say which runtime components an AppImage is missing before its env hygiene by @danielhanchen in https://github.com/unslothai/unsloth/pull/9370
* Drop the HunyuanImage GGUF the Hub no longer serves by @danielhanchen in https://github.com/unslothai/unsloth/pull/9369
* Studio: fix two intermittently failing tests by @danielhanchen in https://github.com/unslothai/unsloth/pull/9200
* Acknowledge the Colab rotation the oracle snapshot is measuring by @danielhanchen in https://github.com/unslothai/unsloth/pull/9376
* Studio: carry a restated instruction once, not eight times by @danielhanchen in https://github.com/unslothai/unsloth/pull/9368
* Stop 13 jobs rebuilding the same frontend on every commit by @danielhanchen in https://github.com/unslothai/unsloth/pull/9375
* Stop 19 test files racing on one PowerShell startup cache by @danielhanchen in https://github.com/unslothai/unsloth/pull/9371
* Stop the pip cache key duplicating itself on every Python patch bump by @danielhanchen in https://github.com/unslothai/unsloth/pull/9377
* Studio: list Ollama models in the chat picker and load their manifest refs by @lonexreb in https://github.com/unslothai/unsloth/pull/9237
* Stop the Studio temp probe leaving anything behind by @danielhanchen in https://github.com/unslothai/unsloth/pull/9298
* Stop the dist cache key hashing 457 files the rebuild check never reads by @danielhanchen in https://github.com/unslothai/unsloth/pull/9380
* Studio: overturn a chat-only MLX verdict the stack contradicts (#9120) by @yzxcj797 in https://github.com/unslothai/unsloth/pull/9124
* Put the managed Node bin dir on PATH when spawning stdio MCP servers by @NilayYadav in https://github.com/unslothai/unsloth/pull/9304
* Studio: normalize replayed tool-call ids for external providers by @NilayYadav in https://github.com/unslothai/unsloth/pull/9116
* Don't select Flash Attention 2 for float32 model loads by @NilayYadav in https://github.com/unslothai/unsloth/pull/9110
* Studio: keep every cached revision's quants on the on-device card by @Lyxot in https://github.com/unslothai/unsloth/pull/9381
* Studio: put Remote and LAN access on their own settings tab by @shimmyshimmer in https://github.com/unslothai/unsloth/pull/9389
* Run six pinned-symbol jobs on one runner instead of six by @danielhanchen in https://github.com/unslothai/unsloth/pull/9385
* Studio: take the model name back off the chat rows by @shimmyshimmer in https://github.com/unslothai/unsloth/pull/9395
* Stop five Windows jobs rebuilding the same frontend on every commit by @danielhanchen in https://github.com/unslothai/unsloth/pull/9387
* Run the Windows UI suites in two parallel lanes instead of five in sequence by @danielhanchen in https://github.com/unslothai/unsloth/pull/9391
* Studio: keep a downloaded model listed when upstream moves on by @oobabooga in https://github.com/unslothai/unsloth/pull/9347
* Studio: offer only the Codex models the ChatGPT plan can actually reach by @NilayYadav in https://github.com/unslothai/unsloth/pull/9302
* Studio: place the vision projector on the CPU only when it does not fit in VRAM by @danielhanchen in https://github.com/unslothai/unsloth/pull/9063
* Fix Qwen3.8 presence penalty defaults by @Imagineer99 in https://github.com/unslothai/unsloth/pull/9372
* Studio: Forward thinking controls and return reasoning on /v1/messages and /v1/responses by @NilayYadav in https://github.com/unslothai/unsloth/pull/9107
* Studio: keep the working llama.cpp install when a prebuilt update can't move it aside by @NilayYadav in https://github.com/unslothai/unsloth/pull/9221
* Studio: add a Disable Vision toggle to a model's Advanced Settings by @danielhanchen in https://github.com/unslothai/unsloth/pull/9383
* Studio: stop the scroll stabilizer invalidating style for the whole thread by @danielhanchen in https://github.com/unslothai/unsloth/pull/9328
* Studio: never offer a calibration imatrix as a GGUF variant by @danielhanchen in https://github.com/unslothai/unsloth/pull/9409
* Stop Codex cancelling its own first turn while llama-server reads the prompt by @danielhanchen in https://github.com/unslothai/unsloth/pull/9403
* Studio: stop a reset dropping the task when the user types short prompts by @danielhanchen in https://github.com/unslothai/unsloth/pull/9379
* Studio: finish a reply that hit Max Tokens instead of asking whether to by @danielhanchen in https://github.com/unslothai/unsloth/pull/9382
* Format the mmproj placement policy tests so pre-commit stops failing every PR by @danielhanchen in https://github.com/unslothai/unsloth/pull/9414
* Studio: scope the sidebar drag's custom-property writes off the thread's ancestors by @danielhanchen in https://github.com/unslothai/unsloth/pull/9400
* Studio: size a large tool result to the window the model is running with by @danielhanchen in https://github.com/unslothai/unsloth/pull/9384
* Report what the model server actually said when a stream fails by @danielhanchen in https://github.com/unslothai/unsloth/pull/9390
* Admit generations against the KV cache, not just the slot count by @danielhanchen in https://github.com/unslothai/unsloth/pull/9392
* Studio: add Mmap/Mlock, draft KV cache dtype, Checkpoints and Cache RAM to Run settings by @danielhanchen in https://github.com/unslothai/unsloth/pull/9410

## New Contributors
* @losuler made their first contribution in https://github.com/unslothai/unsloth/pull/8581
* @yzxcj797 made their first contribution in https://github.com/unslothai/unsloth/pull/8891
* @Eoin-Houstoun made their first contribution in https://github.com/unslothai/unsloth/pull/8464
* @dyuan311 made their first contribution in https://github.com/unslothai/unsloth/pull/9021
* @indrajeetapache made their first contribution in https://github.com/unslothai/unsloth/pull/9002
* @karan68 made their first contribution in https://github.com/unslothai/unsloth/pull/8670
* @sts-change made their first contribution in https://github.com/unslothai/unsloth/pull/8919
* @pascalandr made their first contribution in https://github.com/unslothai/unsloth/pull/9275
* @lonexreb made their first contribution in https://github.com/unslothai/unsloth/pull/9237

**Full Changelog**: https://github.com/unslothai/unsloth/compare/v0.1.800-beta...v0.1.801-beta