News

Qwen-Image 2.1 is on the Media page, at the top of the image list. One 7B model does what the previous Qwen bundle needed two for.

• Text → image. Renders at a native 2K (2048×2048), and the lettering on a sign or a poster comes out spelled right.
• Editing with up to 10 references. Upload the picture you want changed plus anything it should borrow from, then point at them in the prompt as <image1>, <image2> … Faces, products and the rest of the frame stay put.
• Transparent PNG, natively. Ask it to drop the background and you get a real alpha channel — no matting pass, no white halo along the hair line.

Three ready-made workflows ship with it: qwen21-text-to-image, qwen21-image-edit and qwen21-background-removal. Runs from 24 GB.

Open Media → Qwen-Image 2.1.

GLM-5.3-abliterated is in the model list: a frontier-class model with the refusals removed, and fully tested on my platform before release.

• Always ready, no GPU needed. There is nothing to start and no model to wait for. It answers straight away, and nothing is billed while it sits idle.
• Pay per token, not per hour. $3.90 per 1M input tokens, $0.39 per 1M cached input, $6.50 per 1M output.
• Abliterated. It answers instead of lecturing.
• 1M context, with reasoning. Text only.

It runs on a third-party service, which may log conversations.

Pick GLM-5.3-abliterated in the model list, or call imbutus/glm-5.3-abliterated over the API.

AuK is on the Media page: Tencent's 1.5B speech model, driven by plain-language instructions.

• Enhance — the reason it is here. Removes noise and reverberation from a recording of any length, same length out. Any language. Clean a reference clip before you clone it.
• Voice design — a bonus. Describe a voice in words, no reference audio, and AuK speaks the line.
• Also: voice cloning, SRT dubbing composed to fit each cue, re-voicing through a transcript.

Everything that speaks — design, cloning, dubbing — is English and Chinese only; anything else comes back as fluent nonsense. For other languages use Fish Audio S2 or Chatterbox. Runs from 24 GB.

Open Media → AuK → auk-enhance.

Every running bundle on the Media page now hands you root on its machine. Under the ComfyUI login there is a collapsed SSH for pod line: one command to copy, one key to download.

• Fresh key every time. I mint a new ed25519 keypair for each pod and again for each respawn. It is never reused and dies with the pod.
• Root, not a sandbox. Install custom nodes, drop in your own models and LoRAs, read the ComfyUI log, pull finished outputs with scp.
• Nothing to set up. Download the key, paste the command, you are in. The command already does the chmod 600.
• Hand it to your AI agent. Give the key and the command to Claude Code, Codex or any agent with a shell, and it can do anything with the pod for you: install nodes, wire up a workflow, chase an error in the log, fetch the results.
• All bundles. Image, video, voice — every media bundle carries it.

Open Media, start a bundle, and expand SSH for pod once it is ready.

I'm rolling out a new set of agentic security workflows you can run right from chat, live on a rented Kali Linux machine:

• OSINT — email, person, company, domain, phone and username investigations.
• Recon — active scanning: subdomains, ports, services and template-based vulnerability checks, plus a full web-app scan on any HTTP/HTTPS service found — crawling, parameter and JWT decoding, SQL-injection and XSS possibility checks, secret scanning and JWT weak-key detection.
• Attack — turn the flagged possibilities into confirmed vulnerabilities with working, non-destructive proofs of concept, and get a client-ready report with a reproducible attack for each finding.

They work together — each run stores its findings in a shared database on your Kali machine, so later workflows reuse earlier results, and every run saves a downloadable report.

Heads up: these run on a newly rented Kali machine — one created after this post. If you already have an older machine, terminate it and rent a fresh one to get the toolkit.

This is an early announcement: the workflows are new and still in a testing and polishing stage. Try them out and send me feedback through support.

The node has a walkthrough now: 34 minutes covering every documented MiniMax H3 feature, in the order you meet them.

• The timeline. Tracks for the picture, the camera and the sound; the playhead, splitting a block, and why clip length snaps to 5 + 17 frames.
• Entities out of a file. A person, a costume, a prop, a place, a style, a motion, a voice — each becomes a card with a number you paste into any block, and it stays the same across shots.
• Dialogue. Describe a voice, pick who speaks, in what language, on or off screen.
• References. Storyboard, first and last frame, keyframe, and how much of a file survives — including carrying a face from one photo onto a person in another.
• One full example built from scratch and generated at the end.

Watch it: MiniMax Director — the full walkthrough. Dubbed: Русский, 中文.

It is already on your GPU. Start the MiniMax H3 bundle and open the MiniMaxDirector workflow — the bundle loads the newest release from GitHub. On your own machine, install minimax-director from the ComfyUI registry, or clone github.com/imbutus/ComfyUI-MiniMaxDirector into custom_nodes/.

Three models joined the LLMs page and two left it.

• huihui-ai/Huihui-Ornith-1.5-9B-abliterated — small, cheap, and tuned for code. A 9B with image input and 262k context, abliterated. It is also the first model here carrying two labels at once: it is the budget tier and a coding model, so you will find it under either heading.
• huihui-ai/Huihui-Ornith-1.5-35B-A3B-abliterated — the bigger Ornith. Same family, a 35B mixture-of-experts, image input, 262k context, abliterated. It takes over from Ornith 1.0.
• huihui-ai/Huihui-CyberStrike-OffSec-35B-abliterated — a security specialist, in a category of its own. A 35B tuned on penetration testing and red-team work, image input, 262k context, abliterated. Models like it now carry an offsec label instead of sitting among the general-purpose ones.
• Two models retired. The Qwen3.6 35B-A3B and the Ornith 1.0 35B are no longer offered: Ornith 1.5 covers what they did, and the Qwen3.8 27B stays the general pick. Sessions you already ran on them keep showing in your history.
• Labels stack now. A model is no longer limited to a single badge, which is what lets the 9B appear as both cheap and coding.

Open the LLMs page to start any of them.

The node I released in August has had 264 commits since, and 0.16.0 collects them. The idea is unchanged — you lay a film out on a timeline instead of cramming it into one prompt box — but far more of what H3 actually reads now comes out of it.

• Everyone and everything gets a card. A person, a prop, a costume, a place, a look: one card each, holding the picture it came from, its description and its voice. It is the only place a file is described, so nothing is repeated on a block.
• The voices come out of the same pass as the picture. Write the line, pick the face, and it compiles in the exact form H3 was trained on — including a line that crosses a cut, or one cut off by the end of the clip.
• A face can be carried onto somebody else. A card can also take its motion from a video and its timbre from a recording, and each of those is cited inside the subject it belongs to rather than left floating.
• Every file the clip carries sits in one list. Drag one onto a track to place it, or leave it clip-wide for a look that has to hold throughout. A file that has gone missing says so, and the run is refused instead of quietly producing the wrong thing.
• A live linter checks the rules the guides state outright: clip length, reference clips outside 2–15 seconds, a marker that cannot do what it is being asked to do.
• The node is as tall as whatever you have open, and the whole piece exports as one JSON.

Full notes: the 0.16.0 release.

It is already on your GPU. Start the MiniMax H3 bundle and open Docs → Media → MiniMax H3 → the MiniMaxDirector workflow. The bundle loads the newest release from GitHub, so there is nothing for you to update.

On your own machine, clone github.com/imbutus/ComfyUI-MiniMaxDirector into custom_nodes/, or install minimax-director from the ComfyUI registry once 0.16.0 clears review there. Needs ComfyUI 0.31.0 or newer. MIT licensed.

The newest Qwen generation is on the LLMs page: huihui-ai/Huihui-Qwen3.8-27B-abliterated, a dense 27B in full precision, abliterated.

• It takes longer to answer, and the answers are smarter for it. Reasoning runs at maximum effort by default, so the first tokens come noticeably slower than on the smaller models. The wait is the whole point: it works the problem through instead of replying fast. If you want a quick back-and-forth, one of the smaller models is the better pick.
• It reads images. Attach a picture and ask about it, in the same conversation as everything else.
• 262k context. Long enough for a whole codebase or a stack of documents in one session.
• Full precision, nothing thrown away. The weights are the ones the authors published, which is why it needs a large card and about 10 minutes before the first message.

$2.40/hr. Open LLMs and start it.

I wrote a ComfyUI node for MiniMax H3 and released it as open source: MiniMaxDirector. Instead of packing an entire film into one prompt box, you lay it out on a timeline — what happens on screen, how the camera moves, what is heard — and it compiles that into the single structured prompt H3 actually reads.

What it does for you:

• Three tracks. Shots, camera and audio, each segment with its own text and its own span.
• Cut times are computed, so the model is told exactly when each shot begins.
• Legal lengths only. The clip is snapped to a duration H3 accepts, so a run cannot fail on arithmetic.
• Attach a file to a segment and it becomes a numbered reference in the prompt automatically — no tokens to type by hand.

It is already installed on your GPU. Start the MiniMax H3 bundle and open Docs → Media → MiniMax H3 → the MiniMaxDirector workflow; the graph is ready to run, nothing to install.

On your own machine, it is on the ComfyUI registry as minimax-director — searchable in the ComfyUI Manager, or comfy node install minimax-director. Source and issues: github.com/imbutus/ComfyUI-MiniMaxDirector. MIT licensed.