Skip to content
InvokeAI Logo InvokeAI Documentation

All versions since InvokeAI v6.13.5
All versions

InvokeAI v6.13.5

This is a maintenance release of InvokeAI focused on bug fixes and stability. Version 6.14.0 will be the next major feature release, featuring video generation, multiple GPU support, the Wan 2.2 image and video models, pressure-sensitive canvas, HiDiffusion integration, workflow to workflow calls, and much more.

Here are the highlights of the current release:

Updated Libraries

  • React version 19 (frontend support) @joshistoast
  • ROCm version 7.1 (AMD GPU support) @lstein @heathen711
  • Transformers version 5.5.4 (AI inference and training support) @kappacommit

Major Bug Fixes

  • Fix the image viewer progress preview behavior when selecting a gallery image during image generation. @JPPhoto
  • Save and restore both positive and negative prompt history for SD images. @DustyShoe
  • Fix crash when a LoRA was applied to a Flux.2 4B/9B Diffusers model loaded with FP8 storage. @Pfannkuchensack
  • Restore the X/Y queue badges when running in multiuser mode. @lstein
  • Prevent main panel from being squeezed out of existence on small screens. @lstein
  • Gallery maintenance script updated to correctly handle nested image subfolders when structured image directory activated. @lstein
  • Fix crash when using a Z-Image model with regional guidance. @Pfannkuchensack
  • Prevent crash when Heun (2d order) sampler applied to a FLUX.2 Klein model with reference images. @Pfannkuchensack
  • Prevent hang when an unknown wildcard was passed to the dynamic prompt processor. @Pfannkuchensack

New Features

  • Support PEFT named-adapter LoRAs. @Pfannkuchensack
  • Add Qwen Image single-file checkpoint loader with fp8 support. @Pfannkuchensack

Documentation

  • Added information on 3d party GPU providers that host InvokeAI. @lstein @michaelmanly
  • Fix uvicorn URL. @fishd72
  • Add graph execution example to API guide. @wunianze666-netizen

Full List of Changes

New Contributors

Full Changelog: https://github.com/invoke-ai/InvokeAI/compare/v6.13.0…v6.13.5

InvokeAI v6.13.6

This is a patch release.

It corrects a bug introduced in v6.13.5 that caused Qwen Image and Qwen Image Edit models to crash with a dictionary key error when used with the default single-file Qwen 2.2 VL Encoder.

Please see https://github.com/invoke-ai/InvokeAI/releases/tag/v6.13.5 for information on what’s in this release.

Full Changelog: https://github.com/invoke-ai/InvokeAI/compare/v6.13.5…v6.13.6

v6.13.7

InvokeAI Version 6.13.7

⚠️ This is a security patch release

All recent versions of InvokeAI through 6.13.6 contain a security hole that potentially allows an attacker to recover the contents of the running instance’s invokeai.yaml and api_keys.yaml files. These files can contain sensitive data including API keys for image generation hosting services and API download tokens for HuggingFace and other model download sites.

Your keys/tokens may have been compromised if:

  1. Your invokeai.yaml file contains download API tokens for any model provider (e.g. Civitai, HuggingFace),
  2. Your api_keys.yaml file contains API keys for one or more remote image generation services (e.g. Alibaba, OpenAI),
  3. At any point, you set the listen host to 0.0.0.0 to enable network access to your InvokeAI machine, and
  4. Either you exposed the InvokeAI machine to the Internet, or you share your LAN with untrusted individuals.

In the event that these apply to you, or just to be safe, please upgrade to version 6.13.7 and rotate your API tokens and keys.

👉 Note: The InvokeAI team discourages opening your InvokeAI host to the internet. If you choose to do so, we recommend that 1) you use the multiuser configuration and pick strong passwords; and 2) that you isolate the InvokeAI instance in a Docker container. 👈

Additional security fixes

This patch also closes a series of related but less critical security issues:

  • Leakage of model paths: An attacker could recursively enumerate the paths of all models and model-like files, in an arbitrary directory on the Invoke server’s filesystem.
  • Testing existence of arbitrary paths: An attacker could also determine whether a certain file is present on the Invoke server’s filesystem, even if it wasn’t a model file.
  • Administrator functions available to regular users: In multiuser mode, several functions intended for the use of Administrator users were available for use by unprivileged users. Not an obvious security hole, but a potential problem.

Many thanks to George Chen (@geo-chen), for reporting the original issue.

Full Changelog: https://github.com/invoke-ai/InvokeAI/compare/v6.13.6…v6.13.7

v6.13.8

InvokeAI Version 6.13.8

⚠️ This is a security patch release

Recent versions of InvokeAI through 6.13.7 contain two security holes:

  1. An attacker can craft a URL payload that overwrites files in the server’s custom nodes directory, allowing for execution of arbitrary code on the next startup. On Windows systems, the hole could allow an arbitrary file to be placed anywhere on the filesystem that the server has access to.
  2. A second bugs allows the server to be tricked into connecting to arbitrary networks and services on the local LAN or the Internet.

Both scenarios are closed by this release, and InvokeAI users are strongly encouraged to upgrade.

Many thanks to Abhinash Singh and Arpit Jain for independently reporting issue #1.

Full Changelog: https://github.com/invoke-ai/InvokeAI/compare/v6.13.7…v6.13.8

InvokeAI 6.14.0

This is a big release that adds many new user visible features including:

  • Video generation support via Wan 2.2.
  • Krea.2-Turbo and Raw model support
  • Flux.2 Dev support
  • Ernie Turbo model support
  • Ideogram 4 support
  • Anima controlnets and inpainting
  • Flux.2 PiD support (super resolution up to 4K)
  • Multi-GPU support
  • Native Intel XPU support
  • FP8 support for Anima, Flux.2 Krea Turbo & Raw, and Z-Image Turbo

Video Generation

You can now generate short videos using the Wan 2.2 video model. We support text-to-video, image-to-video (use a still image to initiate the video), and image-to-image-video (interpolate video between two images). Video generation is available through the workflow editor, and includes a series of template workflows that allow you to generate videos and concatenate them together into longer productions. See Video Generation to get started.

New models

We now provide support for Krea.2-Turbo, Krea.2-Raw, Ernie Image Turbo, and Ideogram 4.

These are pure text-to-image models (no image editing capabilities). and LoRA/ControlNet/IPAdapter support is being rolled out in stages. Not all LoRAs will currently load. Please report those that don’t in Issues.

Current capabilities are:

ModelText-to-ImageImage-to-ImageInpaintingOutpaintingNegative PromptingReference ImagesRegional GuidanceLoRAsControlNets
Krea-2-Turbo✅✅✅✅⚠️ needs CFG > 1 (off-spec for Turbo)❌⚠️ positive only✅❌
Krea-2-Raw✅✅✅✅✅ (CFG > 1)❌⚠️ positive only✅❌
Ernie-Image-Turbo✅❌❌❌✅ (CFG > 1)❌❌❌❌
Ideogram-4✅❌❌❌❌❌⚠️ prompt + bbox only❌❌
Anima✅✅✅✅✅ (CFG > 1)❌✅ pos + neg✅✅
Flux.2 Dev✅✅✅✅✅ (CFG > 1)✅✅ pos + neg✅✅

Notes

  • Krea-2-Turbo vs Krea-2-Raw — identical feature support; they share one graph builder. The difference is sampling: Turbo is the distilled checkpoint (~8 steps, guidance disabled, fixed mu), Raw is undistilled (~28 steps, CFG ~4.5, dynamic mu).
  • Krea-2 negative prompting — honored only when CFG > 1. The variant does not gate CFG, so the negative prompt is live on Turbo too, but Turbo is meant to run with guidance off; raising CFG to reach the negative prompt is off-spec and degrades output.
  • Krea-2 regional guidance — regional positive text only. Regional negative prompts, auto-negative, and regional reference images raise “unsupported” warnings on canvas. The denoise node supports masked negative conditioning in workflows; the canvas graph does not wire it.
  • Ernie-Image-Turbo — text-to-image only. The denoise node has no denoise_mask input, so masked modes are impossible and image-to-image is not offered; any non-txt2img mode is rejected at graph build. Global negative prompt and the built-in prompt enhancer are supported.
  • Ideogram-4 — text-to-image only; raster layers or inpaint masks with content block generation before enqueue. There is no negative prompt at any CFG: the sampler uses asymmetric CFG with a zeroed unconditional branch, so the denoise node has no negative conditioning input. Regional guidance contributes a positive prompt plus a bounding box to the structured JSON caption; regional negatives, auto-negative, and reference images are dropped with warnings.
  • Anima ControlNets — via kohya-ss’s ControlNet-LLLite adapters (8–66 MB), available as one-click starter installs: sketch (mixed scribble/HED/lineart/grayscale), depth, scribble, lineart, and pose. Control layers work the same way as for other model families. Note that the depth/scribble/lineart/pose adapters were trained on the Preview3 build and are weaker on Anima Base 1.0 — the mixed-conditioning sketch adapter is the strongest general-purpose choice. Each LLLite model may be applied only once per generation. A separate LLLite Inpaint Adapter (Advanced settings) conditions the model on surrounding image content during inpainting/outpainting for cleaner seams.
  • Reference images — unsupported across all five. Anima is explicitly rejected in the canvas validators; the other four have no IP-adapter model config for their base and no IP-adapter wiring in their graph builders.

Multi-GPU Support

If you are lucky enough to have two or more GPUs installed in your system, you can configure InvokeAI to parallelize generation across the GPUs. Two or more queued generation jobs will execute simultaneously on the GPUs, and the preview screen will be split into tiles to show you the progress of each rendering job. This also works when multiple users are logged in: the system will split GPU time fairly among users in a round-robin fashion. If only one generation is queued, then its model’s text encoder will be run on one GPU and the denoiser will run on another, thereby preventing the denoiser from evicting the encoder from VRAM and speeding up the rendering of subsequent images (credits to jacid23 for this concept).

FP8 Support for reduced VRAM usage

You can now reduce your VRAM usage by activating FP8 support for selected models. This reduces the size of models in VRAM by about 50% with minimal loss of quality. To activate it, select the model of interest from the Model Manager, activate the FP8 slider, and Save. The next time you run the model you should see a dramatic decrease in RAM consumption. FP8 storage can be applied to full main models, single-file transformers, and controlnets.

Other Features

  • Performance improvements:

    • We have improved the VRAM consumption estimates for multiple models, which should reduce the number of OOMs.
    • VAEs can now be run on the CPU, reducing the amount of VRAM contention, similar to the text encoders. You can set this option by selecting the VAE in the Model Manager.
    • Improved support for LoRA sliders. We now support UNet-only sliders. In addition, you can now edit the sliders’ min, max and default values from within the Model Manager.
    • If you are using the experimental Wan 2.2 video feature, you can reduce the amount of VRAM used during the VAE decode phase ,at the expense of some performance, by activating tiled decoding. To do so, add wan_memory_optimization: true to your invokeai.yaml configuration file.

  • Canvas improvements:

    • Pressure-sensitive brush opacity on tablets and touch screens.
    • Fine-grid hint added to move tool.
    • A new transparency lock for Gradients and Shapes.
    • Clip Strokes have been extended to the Bbox and Rect/Oval shapes.

  • Workflow management:

    • Graph execution has been optimized, speeding up complex workflows.
    • There is now a LoRA collection picker node which allows you to stack multiple LoRAs together without adding additional nodes.
    • You can now call one workflow from another, making it easier to create and maintain complex workflows.

And lots more! See below for a complete list of What’s New in this release.

Installation and Upgrading

See Installation for various ways to install and run InvokeAI.

What’s Changed Since 6.14.0 Release Candidate 2

New Contributors

Full Changelog: https://github.com/invoke-ai/InvokeAI/compare/v6.14.0-rc2…v6.14.0

What’s Changed Since 6.14 Release Candidate 1

New Contributors

What’s Changed Since 6.13.7

New Contributors

Full Changelog: https://github.com/invoke-ai/InvokeAI/compare/v6.13.7…v6.14.0-rc2

Version 6.14.1 Latest

This release fixes a few bugs in version 6.14 and adds a couple of minor features that didn’t make it into the previous release. For a full rundown of recently-added features please see 6.14.0 Release

New Features in this release

  • Workflows can now be screenshot and saved as PNGs (@JPPhoto)
  • Middle clicks in the gallery will now open the media in a new browser tab. Enable Use Middle Click to Open Images/Videos in New Tab in settings to get this feature. (@DustyShoe )
  • Krea-2 LoRAs built with ai-toolkit’s DoRA magnitudes are now supported (@Pfannkuchensack )

Bug fixes

  • More robust handling of Kohya keys (@Pfannkuchensack )
  • Close a security hole in the handling of AlibabaCloud image downloads (@wunianze666-netizen )
  • Improve VRAM memory reservation for VAEs (@Pfannkuchensack )
  • Improve internal memory management at shutdown time (@lstein )

Full change list with pointers to detailed descriptions:

New Contributors

Full Changelog: https://github.com/invoke-ai/InvokeAI/compare/v6.14.0…v6.14.1

This site was designed and developed by Aether Fox Studio.