All pull requests

Pull requests

• Preserve prompt cache during deferred MCP tool search

  #16546·wasimysaid opened Nov 3, 2026Collaborator·

  ·
  1

• Fix Connections dropdown shrinking on scroll and hover

  #16543·wasimysaid opened Nov 3, 2026Collaborator·

  ·
  1

  1

• Improve model-load error guidance and shorten failure notifications

  #16540·Imagineer99 opened Nov 3, 2026Member·

  ·
  2

• Studio: faster MLX load and reload, ready to serve when loaded

  MLX
  #16536·Lyxot opened Nov 3, 2026Collaborator·

  ·
  11

• Studio: add right-to-left chat layout

  #16534·Aafimalek opened Nov 3, 2026Contributor·

  ·

• fix(studio): reject ambiguous S3 audio references

  important#
  #16529·Max-Reisinger opened Nov 3, 2026Contributor·

  ·
  1

  1

  

• Studio: fit reasoning_effort to a template's enforced allowlist

  codex-converged#edge case#MLXregresses#
  #16527·Lyxot opened Nov 3, 2026Collaborator·

  ·
  12

• Desktop: allow a custom backend port with UNSLOTH_DESKTOP_PORT

  edge case#new-feature#security#
  #16523·danielhanchen opened Nov 3, 2026Member·

  ·
  1

  1

  

• Studio: attach saved Library files to the current chat

  affects UI#new-feature#refactor#
  #16522·danielhanchen opened Nov 3, 2026Member·

  ·
  1

  1

  

• Studio: LiquidAI d1 and Decider decision models, train LFM3 into d1

  codex-converged#MLXnew-feature#
  #16519·Lyxot opened Nov 3, 2026Collaborator·

  ·
  16

• fix(studio): price MLX attention on the route the runtime takes

  edge case#MLX
  #16517·Lyxot opened Nov 3, 2026Collaborator·

  ·
  16

• Studio: size audio.cpp TTS models from measured peaks and run TQ2_1 on the CPU

  edge case#
  #16514·oobabooga opened Nov 3, 2026Member·

  ·
  2

• Unsloth Studio (AMD): retry an unpinned llama-server arch crash on the cards the build has kernels for

  edge case#
  #16513·LeoBorcherding opened Nov 2, 2026Collaborator·

  ·
  4

• Studio: let Laya fine-tunes answer from their GGUF export

  affects UI#edge case#new-feature#refactor#
  #16511·NilayYadav opened Nov 2, 2026Collaborator·

  ·
  

• Studio: keep non-English text in structured Laya instructions

  important#regresses#
  #16510·NilayYadav opened Nov 2, 2026Collaborator·

  ·
  

• Studio: do not download PyTorch Clef for an over-long decision request

  edge case#
  #16509·NilayYadav opened Nov 2, 2026Collaborator·

  ·
  

• Studio: Inception Mercury preset with live diffusion streaming

  affects UI#new-feature#>3000#
  #16508·13770129 opened Nov 2, 2026·

  ·

• Studio: tok/s under replies from API providers

  affects UI#new-feature#
  #16506·13770129 opened Nov 2, 2026·

  ·

• Studio: real token usage and context window for custom connections

  affects UI#new-feature#refactor#>3000#
  #16505·13770129 opened Nov 2, 2026·

  ·

• Studio: use the provider site's icon for custom connections

  affects UI#new-feature#security#>3000#
  #16504·13770129 opened Nov 2, 2026·

  ·

• Studio: bring back the ten Halloween palettes and their composer glow

  affects UI#new-feature#
  #16503·13770129 opened Nov 2, 2026·

  ·

• Studio: add a global token usage limit with a one-time override

  affects UI#edge case#new-feature#security#>3000#
  #16500·debojyoti-dev opened Nov 2, 2026Contributor·

  ·
  1

• Studio: show how to call the Decision API from code

  affects UI#new-feature#>3000#
  #16499·NilayYadav opened Nov 2, 2026Collaborator·

  ·
  1

• Studio: report bad decision datasets before the model downloads

  refactor#
  #16498·NilayYadav opened Nov 2, 2026Collaborator·

  ·
  1

• Studio: open Try after Use in Decision API for Clef and LLM fine-tunes

  affects UI#
  #16497·NilayYadav opened Nov 2, 2026Collaborator·

  ·
  1