All workflows

Workflow runs

• Stop rejecting slow request bodies with a silent 400 (fixes #2)

  completed successfully:
  CI#38:Pull request#8openedbyehartford
  fix/http-request-read-timeout

  Oct 18, 20262m 32s
  More actions for workflow run 38

• Merge pull request #7 from QuixiAI/port/pr6-cpu-engine

  completed successfully:
  CI#37:Commitcf466b0pushedbyehartford
  main

  Oct 18, 20263m 14s
  More actions for workflow run 37

• Port the CPU-Max fork's EmbeddingGemma 2 CPU engine work (PR #6) without the rebrand

  completed successfully:
  CI#36:Pull request#7openedbyehartford
  port/pr6-cpu-engine

  Oct 18, 20263m 15s
  More actions for workflow run 36

• Merge pull request #5 from QuixiAI/metal3-batched-gemma2

  completed successfully:
  CI#35:Commit6b99fe1pushedbyehartford
  main

  Oct 18, 20262m 1s
  More actions for workflow run 35

• EmbeddingGemma 2 on GPUs without Metal 4: simdgroup GEMM and flash attention

  completed successfully:
  CI#34:Pull request#5synchronizebyehartford
  metal3-batched-gemma2

  Oct 18, 20262m 2s
  More actions for workflow run 34

• Fix the new CI jobs against the EmbeddingGemma 2 sources

  completed successfully:
  CI#33:Commit6adc619pushedbyehartford
  main

  Oct 18, 20262m 30s
  More actions for workflow run 33

• Add native Windows CPU support and tested CI binaries

  completed successfully:
  CI#32:Pull request#4synchronizebyehartford
  jethac:main

  Oct 18, 20262m 4s
  More actions for workflow run 32

• Add native Windows CPU support and tested CI binaries

  failed:
  CI#31:Pull request#4synchronizebyehartford
  jethac:main

  Oct 18, 20262m 56s
  More actions for workflow run 31

• Optimize EmbeddingGemma 2 CPU inference with AVX-VNNI, faster attention, and multimodal acceleration

  failed:
  CI#30:Pull request#6openedbyYazeedx0
  Yazeedx0:main

  Oct 16, 20262d 5h 13m 41s
  More actions for workflow run 30

• EmbeddingGemma 2 on GPUs without Metal 4: simdgroup GEMM and flash attention

  completed successfully:
  CI#29:Pull request#5synchronizebyauroter
  metal3-batched-gemma2

  Oct 13, 202626s
  More actions for workflow run 29

• EmbeddingGemma 2 on GPUs without Metal 4: simdgroup GEMM and flash attention

  completed successfully:
  CI#28:Pull request#5openedbyauroter
  metal3-batched-gemma2

  Oct 13, 202636s
  More actions for workflow run 28

• Pack audio towers across requests, speed up media cache hits

  completed successfully:
  CI#27:Commitd8f9e33pushedbyehartford
  main

  Oct 12, 202625s
  More actions for workflow run 27

• Add media tower cache, longest-first wave packing, opt-in persistent …

  completed successfully:
  CI#26:Commitd2a5782pushedbyehartford
  main

  Oct 12, 202651s
  More actions for workflow run 26

• Optimize EmbeddingGemma 2 Metal serving: frame packing, flash tiles, …

  completed successfully:
  CI#25:Commita1b936apushedbyehartford
  main

  Oct 12, 202632s
  More actions for workflow run 25

• Truncate to the context, audit against the model card, add CPU/GPU ex…

  completed successfully:
  CI#24:Commitecc54cdpushedbyehartford
  main

  Oct 10, 202623s
  More actions for workflow run 24

• Let short requests bypass long waves (--short-bypass-tokens)

  completed successfully:
  CI#23:Commit8ddce0bpushedbyehartford
  main

  Oct 10, 202630s
  More actions for workflow run 23

• Add x86-64 kernels to the EmbeddingGemma 2 CPU fast path

  completed successfully:
  CI#22:Commit5488064pushedbyehartford
  main

  Oct 10, 202621s
  More actions for workflow run 22

• Add EmbeddingGemma 2 profile with batched Metal, CPU, and CUDA serving

  completed successfully:
  CI#21:Commit6f4c2e7pushedbyehartford
  main

  Oct 10, 202622s
  More actions for workflow run 21

• Add native Windows CPU support and tested CI binaries

  requires action with the application:
  CI#20:Pull request#4synchronizebyjethac
  jethac:main

  Oct 9, 2026Action required
  More actions for workflow run 20

• Add native Windows CPU support and tested CI binaries

  requires action with the application:
  CI#19:Pull request#4openedbyjethac
  jethac:main

  Oct 4, 2026Action required
  More actions for workflow run 19

• Add linux-arm64 cpu binary to the installer and release process

  completed successfully:
  CI#18:Commit55964e2pushedbyehartford
  main

  Aug 18, 202619s
  More actions for workflow run 18

• Prepare v0.3.1 release

  completed successfully:
  CI#17:Commit19a5da8pushedbyehartford
  main

  Jul 23, 202640s
  More actions for workflow run 17

• Add OpenAI-compatible embeddings endpoint

  completed successfully:
  CI#16:Commit5eefbd5pushedbyehartford
  main

  Jul 23, 202617s
  More actions for workflow run 16

• Add AGENTS.md (CLAUDE.md symlink) with hard-won lessons

  completed successfully:
  CI#15:Commit242e20cpushedbyehartford
  main

  Jul 22, 202616s
  More actions for workflow run 15

• Document the 2x investigation and its opt-in backend flags

  completed successfully:
  CI#14:Commite3ab791pushedbyehartford
  main

  Jul 22, 202617s
  More actions for workflow run 14