All pull requests

Pull requests

• mlxrunner/model/lfm3: release temporary speech cache views

  #19438·pd95 opened Oct 31, 2026Contributor·

  ·
  1

• Add tokencut (context compression proxy) to Infrastructure & Deployment

  #19434·attnstack opened Oct 31, 2026·

  ·

• mlx: load bundled CUDA libraries in-process on Linux

  #19433·dhiltgen opened Oct 31, 2026Collaborator·

  ·

• parser: normalize CR line endings inside quoted Modelfile values

  #19431·mirzasnowy opened Oct 31, 2026·

  ·
  1

• README: add SuperEditr to community integrations

  #19429·sghael opened Oct 31, 2026·

  ·

• tools: resolve JSON function names from call envelopes

  #19426·shoutoutuoadi325 opened Oct 31, 2026·

  ·
  1

  2

• parser: accept case-insensitive parameter names and message roles

  #19423·sunhouy opened Oct 31, 2026·

  ·
  1

  1

• mlxrunner: abort generation when a token repeats past the limit

  #19422·heyemlee opened Oct 31, 2026·

  ·

• manifest: check the digest before a blob reaches the store

  #19421·sunhouy opened Oct 31, 2026·

  ·
  1

  2

• server: give MLX loads the llama-server memory pre-flight

  #19419·nikolowry opened Oct 31, 2026·

  ·
  1

• anthropic: keep all thinking blocks of an assistant message

  #19416·Shayan12456 opened Oct 30, 2026·

  ·
  1

  1

• macos: best-effort support for Intel Macs on macOS 12 (Monterey)

  #19413·1-DARK opened Oct 30, 2026·

  ·
  1

  1

• README: add schemagate to community integrations

  #19412·ashishsinha1602 opened Oct 30, 2026·

  ·

• README: (Community Integration) Add BashBrunch

  #19411·AttAditya opened Oct 30, 2026·

  ·

• cmd/launch: use thinking metadata for OpenCode settings

  #19409·hoyyeva opened Oct 30, 2026Contributor·

  ·
  3

• Fix long System One requests for MLX decision models

  #19408·dhiltgen opened Oct 30, 2026Collaborator·

  ·
  3

• README: add Wavefunction to community integrations

  #19406·stoyan-stoyanov opened Oct 30, 2026·

  ·

• Add Lirnyk to community integrations

  #19402·cheyako opened Oct 30, 2026·

  ·

• openai: make compatibility response IDs unique

  #19401·phanipaladugula opened Oct 30, 2026·

  ·
  1

  3

• model: add MiMo V2.6 model support

  #19400·dhiltgen opened Oct 30, 2026Collaborator·

  ·

• Add AutoTrust decision model to MLX

  #19399·dhiltgen opened Oct 30, 2026Collaborator·

  ·

• server: reject clips the model has no encoder for

  #19397·sunhouy opened Oct 29, 2026·

  ·
  1

  2

• README: add Clairvoyance to community integrations

  #19395·HatcheData opened Oct 29, 2026·

  ·

• README: add RedAmon to community integrations

  #19394·samugit83 opened Oct 29, 2026·

  ·

• llm: don't count CPU-resident input tensors in the VRAM prediction

  #19393·KafKafrnZ opened Oct 29, 2026·

  ·
  1