Issues

• [bug]: [v7.0.0a] Wildcards do not cycle/shuffle on invoke

  bug
  invoke-ai/InvokeAI#9698 · Filter by author adoermenen opened on Oct 7, 2026

• Follow up on graph collector validation findings from #9686

  invoke-ai/InvokeAI#9695 · Filter by author JPPhoto opened on Oct 7, 2026
  

• [bug]: Server error when generating images with Kreation v3 checkpoint for Krea 2

  bug
  invoke-ai/InvokeAI#9691 · Filter by author jackofallbrakes opened on Oct 6, 2026
  3

• [bug]: Expand Prompt popover closes mid-run and can't be reopened until the expansion finishes

  invoke-ai/InvokeAI#9690 · Filter by author unmarco opened on Oct 6, 2026

• External model sizing: Gemini 3 21:9 presets always rejected, snapped size can exceed max, Grok/DALL-E ratios unselectable in webv1

  bug
  invoke-ai/InvokeAI#9680 · Filter by author lstein opened on Oct 4, 2026

• v7 delete confirmation has no "Don't ask me again" option

  bugfrontend
  invoke-ai/InvokeAI#9676 · Filter by author lstein opened on Oct 4, 2026
  1

• [bug]: Prevent stale in-flight model loads from reentering cache

  7.0.0bug
  invoke-ai/InvokeAI#9674 · Filter by author JPPhoto opened on Oct 4, 2026
  

• Multi-user mode: v7 has no controls to share boards or workflows

  7.0.0bugfrontend
  invoke-ai/InvokeAI#9673 · Filter by author lstein opened on Oct 4, 2026

• FLUX.2 starter bundle installs gated, non-commercial Klein 9B and FLUX.2 [dev]

  bugmodel manager
  invoke-ai/InvokeAI#9670 · Filter by author lstein opened on Oct 4, 2026
  

• Model cache: evicting a VRAM-resident model from RAM leaves its weights pinned in VRAM while a node still references it

  7.0.0
  invoke-ai/InvokeAI#9667 · Filter by author lstein opened on Oct 3, 2026
  1

• [bug]: Floating-window movement persists viewport clamp on untouched axis

  bug
  invoke-ai/InvokeAI#9650 · Filter by author JPPhoto opened on Oct 2, 2026
  

• [enhancement]: Support multiresolution inspection of very large images

  invoke-ai/InvokeAI#9615 · Filter by author kvnloo opened on Oct 1, 2026
  1

• Canvas Shape tool creates a raster layer instead of filling the selected inpaint mask or regional guidance

  bugfrontend
  invoke-ai/InvokeAI#9616 · Filter by author lstein opened on Sep 30, 2026

• Canvas: layer settings (regional prompts, control model) are hidden below the tool settings

  bugfrontend
  invoke-ai/InvokeAI#9617 · Filter by author lstein opened on Sep 30, 2026
  1

• Gallery deletion does not check whether media is still in use

  bugfrontend
  invoke-ai/InvokeAI#9618 · Filter by author lstein opened on Sep 30, 2026

• Possible model-support issues found while documenting model families

  bug
  invoke-ai/InvokeAI#9619 · Filter by author lstein opened on Sep 30, 2026
  2

• [bug]: Z-Image turbo regional guidance.

  bug
  invoke-ai/InvokeAI#9612 · Filter by author Jokerlorian opened on Sep 30, 2026

• FP8 Storage is switched on at install for MXFP8 checkpoints, where it costs quality

  invoke-ai/InvokeAI#9620 · Filter by author Pfannkuchensack opened on Sep 30, 2026

• [enhancement]: Upscaling

  enhancement
  invoke-ai/InvokeAI#9608 · Filter by author valenok-polny opened on Sep 29, 2026
  4

• [bug]: Align Graph.add_edge and validate_self collector type validation

  bug
  invoke-ai/InvokeAI#9610 · Filter by author JPPhoto opened on Sep 26, 2026
  

• [enhancement]: Dynamic Tag Filtering for Images and Videos

  enhancement
  invoke-ai/InvokeAI#9601 · Filter by author luisfilipeac91 opened on Sep 26, 2026

• [bug]: 6.14.1 regression with pytorch_cuda_alloc_conf: backend:cudaMallocAsync — Z-Image bf16 + LoRA takes ~30 min whenever the transformer is (re)loaded (VRAM overflows into Windows shared memory)

  invoke-ai/InvokeAI#9597 · Filter by author cochbild opened on Sep 24, 2026
  2

  

• [enhancement]: Upgrade python from 3.12 to 3.14

  enhancement
  invoke-ai/InvokeAI#9621 · Filter by author aleyan opened on Sep 23, 2026

• [enhancement]: add Qwen-Image-2.1 support

  enhancement
  invoke-ai/InvokeAI#9594 · Filter by author krakotay opened on Sep 21, 2026
  2

• [bug]: Error generating simple image on NVIDIA using FLUX.2 Klein 4B

  bug
  invoke-ai/InvokeAI#9593 · Filter by author gafda opened on Sep 19, 2026
  14