Skip to content

v1.60.1 (Cloud Servers, Generation Presets, Full Window, Faster Scan Page)

Choose a tag to compare

@DemonGatanjieu DemonGatanjieu released this 23 Oct 06:01
a6115df

🔍 Scanning (扫描)

  • The scan page opens at once, also with many models: it shows the last counts right away, says it is updating them, and shows the fresh ones when they come. What each model file's header says is now kept on disk (user/anomalous/model_heads.json, by file, read again only when the file changes), so a big library or a network drive is not read again after every restart either; requests that come while the page counts share that one count.
  • See the models found on Civitai: the scan page's "Found on Civitai" count opens a list of them, with each model's Civitai name and base model, a box to find one by name, and a way to open each.
  • Before an online scan, the plugin checks whether Civitai can be reached, on both its addresses. When it cannot, it asks instead of waiting on every model: Scan from the files (base models found, no covers or trigger words; a later scan fills them in), Set a proxy (opens Settings → Models → Network), or Cancel. A check that reached Civitai is trusted for five minutes.
  • Civitai's other address (civitai.red) when civitai.com does not answer: some networks reach only the second, a cloud server in mainland China among them (tested on AutoDL: civitai.com timed out, civitai.red answered in about a second). Scans, import and Model Check's lookups and Settings → Network → Test ask civitai.red then, so models still get their names, trigger words and base models; the test says which address answered. The covers are on Civitai's picture server, which such networks often cannot reach either: after two covers in a row that would not download, a scan fetches no more covers instead of waiting on each.

🧩 Combos (搭配)

  • A combo's models are cover cards, as in the model library: the cover on top, the name below. In a combo saved from picked nodes, the slot (Main model, LoRA…) is written on the cover and a click on the card changes the model; in a default combo, the main models and LoRAs to pick from are cover cards side by side. They used to be rows as wide as the window, with Change far from the name (模型卡片不再横向拉满整个窗口).

🎙️ Character voices (角色语音)

  • Update Anomalous_TTS from GPT-SoVITS Settings (在 GPT-SoVITS 设置里更新 Anomalous_TTS): a new Anomalous_TTS version line shows the installed version; Check for updates looks up its published stable releases (only then does it go online), Update to v… switches to the newest one and offers to restart ComfyUI, and Back to v… undoes it. Anomalous_TTS now publishes numbered stable releases (the first is v1.0.0), so an update never lands on unfinished work. Only an Anomalous_TTS installed with git from its GitHub repository is changed (as ComfyUI Manager installs it), only from the computer ComfyUI runs on, and never over files you changed. When an import needs a newer Anomalous_TTS, the message points here.

  • Generation presets (生成预设): GPT-SoVITS Settings has three presets, Steady (fewer slips, flatter), Standard (the old default) and Lively (livelier, slips now and then): pick the default one for every character, and one for each pair of languages, such as a Japanese character speaking Chinese (until now every such line was made steadier without asking; that is the rule you start with). With nothing changed, voices sound exactly as before.

  • Presets you can hear and compare (听得到、比得出的预设): pick a preset and see where it sits between steadier and livelier next to your default (a ring for the default, a dot for the one picked), with a sentence such as "a little livelier than your default (Standard)". ▶ Listen to this one has a character read a line with that preset; every take uses the same seed, so only the preset changes, and the presets already heard stay as buttons to click back and forth. Folded under it are the numbers as bars against your default, the way the Current node card shows a change (old → new, ↑ ↓); each has its plain name and the usual one beside it (top_k, top_p, temperature, repetition_penalty), so values from the GPT-SoVITS WebUI are easy to match. Click a number to change it; a changed built-in number shows its built-in value, and Reset puts it back. Make default sets the preset every character uses. + New makes your own preset from a copy of the one picked, with a name of your own (such as 念旁白 for narration); it can be picked everywhere the built-in ones can, renamed, and deleted (the question says what uses it; those places go back to the general setting).

  • Modes on Voice-over (配音时直接选模式): under How it is read, Auto (the presets and the character's own settings decide, line by line) or any preset for the whole script, shown against Auto the same way and heard from the ▶ on each line: it reads that line in the mode picked, and the modes already heard on it stay on the side as buttons to compare. The scale and the numbers fold away there, so the narrow side stays short. A preset's numbers can be changed right there and are saved into that preset. Links lead to the presets and to the character's generation settings. This replaces the folded fine-tuning numbers, which showed the node's defaults rather than what the presets used; Save as default now keeps the language and speed only.

  • A character's own generation settings (角色特调): a new Generation settings button on each character card sets the preset and speed for the character as a whole and for each language it speaks, with the numbers themselves when a preset is not enough. Under each row are the numbers its lines will get, blue where they differ from the general presets. The node follows these settings too, so a workflow in ComfyUI sounds like generating here; numbers changed on the node itself still go first. Needs Anomalous_TTS v1.1.0 (update it from the same dialog).

  • More room on Voice-over (配音页更宽敞): each line is one row (number, text, its buttons), so more lines fit; a bar between the lines and the generating side can be dragged to share the width (remembered; double-click for the default, about half each); Into your own workflow is folded to one line; who reads and their voices share one row, and the how-to line shows only until there are lines. When the side is wide enough, the modes' scale and numbers are shown open.

🪟 The window (窗口)

  • Full window (全屏): a new button beside Dock makes the plugin fill the whole ComfyUI page, and turns it back; a double-click on the title bar does the same. It is remembered.

☁️ ComfyUI on a cloud server (云服务器上使用)

  • How far it goes: fully supported on overseas cloud servers (RunPod…). On servers in mainland China (AutoDL…) only in part: browsing, offline scans and tidying work, but Civitai is mostly out of reach there, so lookups are unreliable and covers rarely download (海外服务器完整支持;国内服务器只能部分支持). Not fully tested yet: one AutoDL server so far, no overseas one. The cloud guide ends with a section for AI assistants (what is going on, the steps, how to check, what to do when it did not work, what to tell the user about safety), so a user can hand the link to their AI.
  • On a rented cloud server (AutoDL, RunPod…) the page is always opened from another device, so since v1.58.1 the plugin only let it look: scans, downloads, moves and saves were refused. Now there are two ways: open ComfyUI through an SSH tunnel (the address starts with 127.0.0.1; it counts as the computer ComfyUI runs on, nothing to change), or make an empty file named allow_remote_changes in the plugin folder on the server and restart ComfyUI. With the file, every device that can open the page may change things; the page can neither make nor remove it, and the MCP endpoint stays for this computer only. How, and the risk: docs/guides/cloud-server.md.
  • The View only chip beside the page title can be pressed: it says what is refused and the two ways above, with a Copy button for the one line to paste into the server's terminal (the plugin's real path on that server already filled in), so there is no folder to find. With the file there it reads Remote changes on, in yellow, and its dialog gives the line that turns it off again.
  • Models that are links (模型是快捷方式/链接时): on a cloud server such as AutoDL every model in the models folder is a link into the platform's shared model disk. Scans wrote the info beside the link, but the cards read it from the disk behind it: a scanned model showed as not scanned, under the disk's file name (6b4679…) instead of its own, and editing, deleting, opening its folder or finding it from a saved workflow could be refused as "outside the models folder". A linked model is now taken as named, with its covers and info beside the link; edits, renames and deletes act on the link, never on the shared disk; moving one in Tidy moves the link instead of copying the model disk's file onto the server. Names leaving the folder (..) are refused as before.
  • Models linked from another tool keep their info: a model that is a link with no info beside it shows the info beside the file it links to (A1111's Civitai Helper leaves .civitai.info there). Only read there; covers, edits and scans stay beside the link.

v1.60.0 (Model Import, Tidy, Parameter Compare, Workflows Rebuilt, Base Model Filter, Settings Tabs) — 2026-10-22

📥 Drop a model in, it goes to the right folder (拖进来自动归位)

  • Models page → + Import, or drop model files anywhere on the models page: each file gets a card saying what it is — Checkpoint, LoRA, VAE, text encoder, diffusion model, ControlNet, Embedding, CLIP Vision, upscale model — with its base model, how sure that is, and one line on what such a model is for. It is read from the file's header only; .ckpt/.pt files, which can carry code, are never opened (you pick their type).
  • Where it will go is shown and can be changed; by default it follows the download setting (by base model, such as loras/Illustrious). A file found in your Downloads folder or on the Desktop is moved there (instant on the same drive; tick "Keep the original" to copy); any other file is sent as a copy.
  • While you look, the whole file is checked against Civitai: SDXL-family LoRAs get their exact base model (Pony, Illustrious, NoobAI cannot be told apart from the file alone) and the model and version name. A file you already have (identical content) is pointed out and left out unless you tick it; a different file of the same name is saved as "name (2)". Nothing is ever replaced. The new files are scanned right after.

🧹 Tidy: models in the wrong folder, files you have twice (整理模型)

  • The models page has tabs: All models, Tidy and Sources (Sources was a chip among the model types). Tidy shows a count when it found something.
  • Tidy reads the "table of contents" inside every model file and shows, as cards with the model's cover and name, two steps. 1. In the wrong folder: what the file really is and where it belongs, drawn as Checkpoint → UNet / Anima, with one plain reason (for example an Anima, Flux or Wan model that is only the main part, which "Load Checkpoint" cannot read — Civitai labels all of these "Checkpoint"). Move there takes the model with its cover, info and your notes, keeping its subfolder ("Change place" to pick another); nothing is replaced, and it says which loader node your workflows need instead. 2. Identical files: the copies side by side, the one in the right folder marked to keep (click another cover to keep that one), and one button sends the extra copies to the Recycle Bin with their covers and info. Scanned models are compared at once; same-size files never scanned when you press Compare.
  • Models that load where they are anyway (a whole checkpoint in the UNet folder) are not listed.
  • Not used lately (很久没用的模型和节点包), under the two steps, when you press Look for them, for models under Tidy and for node packs in Settings → Node packs (a line under Tidy leads there), from not used for 30 days, for 90 days to never used. AMB now notes what each run uses and what each workflow you open holds (nodes and models, kept on this computer), and counts the tabs open when you look, so a node pack making sounds or videos, a workflow never saved, or a node only put on the canvas counts too; before that note began, your images and sounds (each records its models and nodes) and your saved workflows, recipes and combos tell. Nothing that came in the last 30 days is judged, and while the note is younger than the group you look at, the page says so.
    • Models, biggest first, with the day they were last used and a warning when a saved workflow, recipe or combo still names one. Tick the ones to drop and send them to the Recycle Bin with their covers and info.
    • Node packs with how many nodes and how long installed (from the day its folder was made), and a warning when another pack's nodes take what only this one makes. Packs that work only through nodes come first; those that also add buttons, menus or panels to the page are listed apart, since their nodes going unused does not say the rest is. A pack used only through its panels or buttons leaves no record: press I still use it and it is listed apart from then on (Unmark undoes it). Find it in the Manager opens ComfyUI Manager's node list already searched for that pack (its name is also copied), where you disable or remove it. Packs without nodes, which only work on the page or the server, are listed apart: no record can tell whether they are used.
  • Importing, uploading and Tidy's moves only work on the computer ComfyUI runs on; opened from another device (ComfyUI started with --listen), they say so instead of writing into your model folders.

🧬 Filter models by base model (按底模筛选)

  • Under the type chips, one chip per base model the listed models have, with counts: All · Illustrious 21 · Anima 7 · SDXL 9 · Unknown 2. Spellings of one family count as one (SDXL 1.0 and SDXL). Models not scanned yet are listed by what their file's header tells. Handy for LoRAs: pick the base model of your checkpoint and only the LoRAs that fit are left. It works together with the search.
  • The scanner's offline guess of a base model (for models Civitai does not know) now also recognizes LoRAs, Qwen-Image, Wan, Chroma and Hunyuan Video files.

✏️ A new model editor, with a source link (模型编辑页重做,能填来源链接)

  • Edit on a model opens a cleaner editor: the cover large on the left with the choices under it (the current one, Civitai's own cover when you replaced it, your newest pictures, upload, or the whole gallery), and on the right Name, Source link and Notes, each saying what it does. Nothing is written until Save; Esc or a click outside cancels. Renaming the file on disk is a tick under the name that says what it means.
  • Source link (来源链接): where you downloaded a model (LiblibAI, HuggingFace…), its site named as you type. The details get a button that opens it, beside Civitai's; shared workflows and Model Check point to it when someone misses that model. Only web addresses are kept.
  • Reset clears the name and notes you set and brings Civitai's cover back; the source link stays.
  • Saving no longer marks the name or notes as yours when you did not touch them.

🪪 Model details, rearranged (模型详情页重排)

  • From a Bilibili comment: the buttons were small and spread over five places, and the file name showed twice. Now the title bar is gone and everything sits beside the cover: a small ← Models / LoRA to go back (Esc does the same; from Model Check or the assistant it says where it leads), the model's name, the size / base model / source line with the Civitai and source links, and one row of buttons: Add node (the one coloured button), Edit, Folder (shows the file in Explorer; pointing at it shows the path) and Outputs (pictures made with it).
  • The file's full path shows once, at the foot of the details with Copy and Delete model. Delete is no longer a red button next to Add node; it asks first in AMB's own dialog, and the files still go to the Recycle Bin. The ⬇ (jump to the bottom) button is gone.
  • In a narrow window (the browser docked beside the canvas) the cover column is hidden as before, so a small cover now sits beside the name.
  • A model's description (Civitai's own text) keeps its headings a little above the text instead of poster size, and a prompt quoted in it wraps instead of making the details scroll sideways.

🗂️ The folder column follows the type; open folders in Explorer (文件夹栏跟着类型走,一键打开文件夹)

  • The column left of the models used to list every type's folders (raw names like checkpoints (1), [Root] rows with 0, nothing marking where you were), repeating the chips above it. Now it shows the folders of the type you picked: All, then its subfolders as a tree whose numbers include their subfolders (empty ones dimmed, ▾ folds one), and In the top folder for models lying directly in the type's folder. The folder you are looking at is marked.
  • 📂 Open next to the type's name opens that type's folder in Explorer, and a 📂 on any folder you point at opens that one, so you can drop files in or tidy up by hand. Like the details' Folder button, it works only on the computer ComfyUI runs on.
  • In a narrow window (docked beside the canvas) the column opens as a drawer over the page; closing it no longer squeezes the page for a moment as it slides away.

🎨 The Combos page looks like the others (搭配页和其他页面统一)

  • It no longer sits in a lighter, inset box with a shadow, a second "Combos" title and a red ✕: it fills the page in the same colours as Models and Workflows. The explanation of what a combo is folds behind ? (shown by itself while you have no combos), and the editor's Back reads ← All combos.

🧩 Combos with rgthree's and Easy-Use's LoRA nodes (搭配支持常用的第三方 LoRA 节点)

  • Stacking many LoRAs used to mean a long chain of Load LoRA nodes. New on the Combos page now also offers Default combo, LoRAs in Power Lora Loader (rgthree: every LoRA in one node, each row with its own switch and strength) and …in Easy-Use LoRA stack (up to ten per stack node, put on the model by Easy Apply LoraStack), each only when that node pack is installed. You still just pick the main model and the LoRAs; the combo's editor can switch between them later.
  • A combo set to a node pack this ComfyUI does not have is put down with ComfyUI's own LoRA loaders, and says so.
  • Save as combo on canvas nodes keeps a Power Lora Loader's LoRAs: before, its rows were lost and it came back empty. Each row's LoRA is now a slot you can change in the editor, its strength and switch shown beside it.
  • The LoRAs inside a Power Lora Loader are seen everywhere else too (Power Lora Loader 里的 LoRA 也认得了): before, AMB looked only at ordinary drop-downs, so a downloaded workflow using one never said its LoRAs were missing. Now Model Check lists each switched-on row's LoRA (missing, found by fingerprint, download, put back in), saving a workflow records their fingerprints, workflow cards and a picture's parameters list them with their strengths, a workflow's models used include them, and a LoRA's Outputs finds the pictures made with it. A workflow's models used also include Easy-Use LoRA stack rows. Only these widely used nodes are read this way; ComfyUI's own loaders remain the reference.

🚀 First-run setup on Home (新手一键准备)

  • Until your models have been scanned once, Home opens with New here? Let's get set up: four steps and Set up for me. It checks that Civitai answers (many need a proxy for it; if not, Retry or Scan offline), offers the optional Civitai key, scans every model with a progress bar in the card (it renames no files and runs in the background), then starts the 30-second tour. It stops to ask when the network is down. Later hides it; Home's New here? line has Set up to bring it back. It is a card on Home, never a popup.

≈ What was read from the file says so (离线推断标清楚)

  • A model Civitai does not know has its base model inferred from the file, as before; now it says so. Its card reads 推断 SDXL (it was ≈), and its details open with a note: why Civitai did not have it, that the base model was read from the file and may be off, and Add a name and source link, instead of an English placeholder description.
  • The scan page says it too: an online scan "looks up 5 on Civitai and infers the base model from the file for any it does not find", and the state chip reads Not found · inferred from file.

🏷️ Version labels on model cards (卡片显示版本号)

  • A scanned model's card shows its Civitai version ("v2.0", "Illustrious", "fp8"…) as a small label above the name, so versions of one model that share a file name can be told apart.

⇄ Compare parameters (参数对比)

  • Two pictures from the same workflow often differ in only a few values. Press ⇄ on one picture in the Gallery, then on another (or Compare with the canvas); the picture details have the same ⇄ for the canvas. The window says whether both use the same nodes and lists only what differs, node by node, A and B each in its own colour: prompts by the tags only A or only B has (marked positive or negative), long names one line per side with the differing part marked, seeds folded away, nodes only one side has listed apart. Two workflows numbered differently still compare the nodes each has one of, such as their only KSampler.
  • Against the canvas, tick the values to take from the picture and put them on the canvas in one go (seeds stay); one Ctrl+Z or the toast's Undo takes them back.
  • What changed since the last picture (和最新出图对比): ⇄ Canvas vs. newest picture at the top of the Gallery and of Activity compares the canvas with the newest picture in the output folder, edits made by hand on the canvas included, and can put the picture's values back.
  • Key parameters (关键参数): press ☆ before a value's name (tick Show equal values too to star one that matches) and it stays at the top of every comparison, for every workflow, kept with your ComfyUI settings.
  • Current node shows where saved values came from (当前节点的参数方案:出处、出图、并排): each saved set for the selected node is a card with the picture made with those values (or the picture they were kept from; click to see it large) and a line on where it came from, with the day: a parameter set of workflow X (with how many pictures it made, and Open to show that set in the workflow), kept with picture Y (See it), or saved here. The values it would change are drawn instead of written: numbers as two bars (now and then) with ↑ or ↓, choices as the old one struck through → the new one. What stays (seeds, models, choices this computer lacks) is said with why; Show all lists every value, equal ones too (there is no more "4 shown, N more"). ⇄ Compare opens the comparison of the whole workflow with the canvas, not only this node.
  • Where the node's values came from (上次套用的是哪套): a line on top of the parameters says which set was last put into this node and when ("上次套用「阿罗娜K4.0 · 02」· 今天 18:02"), and what changed since ("之后变了:cfg 5 → 6.5"), with Put them back; that set's card is marked. It is remembered in this browser, per workflow and node, and forgotten when the change is undone. Without it, the line says which saved set the values match, or come closest to ("只差:cfg 6 → 5").
  • Side by side (并排): with two sets or more, switch to a table: the current values and each set in columns (its picture on top), the values a set would change marked, Use N under each. The choice is remembered.
  • Parameter sets in view (参数方案更好找): a workflow card with saved sets shows ▤ N (how many), which opens them directly. Each set there shows its key values (those unlike the workflow's own in colour), how many values differ, and ⇄ to compare it with the workflow.

🗂️ A tidier Workflows page (工作流页面重做)

  • One bar on top: search, Filter (kind and tag in one menu), grid or list, share and import, and Save the canvas. The drag tip strip and the red ✕ are gone (Esc still closes the page; the empty page explains dragging).
  • A workflow opens on its values (打开工作流先看参数方案): the details are rebuilt around them. On the left: 默认参数 (the workflow's own), your saved sets (each with its key values and how many values differ), + 从画布保存 and + 新建参数方案, then 全部出图 and 版本. On the right, the chosen set: 应用到当前画布, ⇄ 对比 (with the default values or the canvas), ⋯ (copy as a new set, rename, delete), its main values, its prompts (long ones folded, the role still chosen from a menu), the pictures made with these values, and, folded, the models and every node's values. The bar keeps the name (click to rename), the models' state, Open and ⋯ (edit, export). The old Overview tab is gone: cover, notes and tags sit at the top of 默认参数.
  • A new set is named after the workflow with a number (阿罗娜K4.0 2); values saved without names show their real names instead of 控件 1, 控件 2. A set equal to the workflow's own values (saving a workflow keeps one) is no longer listed twice.
  • Compare pictures inside a workflow too: every picture in 全部出图 and under a set has ⇄ — pick two, or one and the canvas or the default values. The pictures there now load as thumbnails.
  • Cards keep the picture, the name, a line of steps, sampler, LoRAs and size, and the model. The models' state is a dot on the cover, with words only when some are missing; 功能模块 marks a partial workflow. ▶ Open and ⋯ (edit, export, delete) are two small round buttons in the cover's corner, shown when the pointer is on a card, so pressing a card to look inside no longer opens it on the canvas by mistake.
  • Focus values (重点参数): mark what matters most in a workflow. Under 全部节点参数 every value has a ☆; marked values show at the top of 默认参数 and of every set as large tiles with the node they belong to, a note you can add (「0.6–0.8 最好」), and, on a set, the default value when it differs. Set rows list them first, and in ⇄ 对比 they lead the key values with their note. A star turns at once with a little bounce, the value fades in at the top, and marking or noting one saves only that (no new version of the workflow). They belong to that workflow (the ☆ in the comparison window still stars a value for every workflow), are kept when the workflow is updated from the canvas, and travel with it in an exported workflow package (share codes carry only the canvas workflow).
  • Tips where they help (参数方案的引导): a workflow without sets says what they are for, with + 从画布保存 right there; on a set, a line points at ⇄ 对比 until you have used it (or pressed 知道了).
  • Folds remember you (折叠记住你的选择): 用到的模型 and 全部节点参数 open as you last left them. 用到的模型 says on its title whether the models are all here or how many are missing, and opens by itself when some are.
  • Docked, the details fit (侧边栏模式不再挤): in a narrow panel the left column becomes one row of pills above the values (默认参数, your sets, new, 全部出图, 版本).
  • Saving a workflow, redone (保存工作流的窗口重做): the cover is large on the left, chosen from the canvas, the current cover or your recent pictures (as thumbnails); name, tags and notes on the right; Cancel and Save always in view. The name starts as the canvas's workflow name when ComfyUI has saved it under one. Esc cancels, Enter in the name saves.

🔇 Models you never want scanned (不需要扫描的模型)

  • A model whose header looks like no drawing model (a voice model such as F5-TTS kept in checkpoints, say) is no longer just "not scanned" forever. The scan page asks about them: 都不扫, 都照常扫, or 逐个看. Set to not be scanned, a model shows 不扫描 instead of 未扫描, a scan of everything skips it (picked by hand it is still scanned), and the scan page lists these models to change it back. A model's details have the same button. Nothing is decided for you: scanning still keeps a fingerprint that repairs workflows.
  • Not on Civitai is no fault (Civitai 上没有不再像出错): the count of models Civitai did not find warns only when some file is damaged or a lookup is still to do; each such model says its fingerprint is kept for repairing workflows.

🎯 The scan page, rearranged: pick models beside the button, settings in place, components apart (扫描页重排)

  • Quieter the further down. On top one card answers "what now": two rows of your models' covers (an empty dashed frame for each model not scanned yet, gently pulsing while a scan runs), what is left in one line ("3 models still to look up"), the scan button, Pick models to scan…, and Online / Offline right beside them. Then the models' state as a coloured bar with a chip per state; then, only when there is something, what needs a look (damaged files, voice-like models, models set to never be scanned, other formats), one row each with its icon; then the settings as tiles with an icon and a switch each; the last scan is one line at the foot. "How scanning works" is folded under the title.
  • Pick models to scan… sits beside the scan button. The picks are named under it (each with a ✕, with their count and total size) and the button scans just them; Clear, scan everything again goes back. There is no "all / only ones I pick" switch any more. When some picks are already found, it says so: a scan only reads their record again unless Fetch again is on.
  • Under the button, what this scan will do, in one sentence; Details unfolds the numbers: how many will be looked up on Civitai (not scanned yet, and those last scanned without a connection; how many of them are components), how many are already found and keep their record, how many Civitai does not have and are not looked up again, how many are set to never be scanned. Turning on offline or Fetch again changes the lines at once.
  • The scan settings are on the page instead of a page of their own: Civitai names on cards, fixing the workflow afterwards and the Civitai key as tiles; renaming files and fetching everything again are folded under More, marked as changing files, and when one is on the scan card says so in yellow.
  • Components on their own (基础组件): VAE, text encoders, CLIP, CLIP Vision and upscalers get a chip of their own, "7 components · not found is fine", whose list (found, not found, not scanned yet) says that many of them are not on Civitai at all and that this does not matter (the fingerprint is still kept for Model Check). The counts above, and the Not found and Not scanned yet lists, leave them out; their own list shows all of them.
  • The model picker is new: every model a scan reads, as a list (name, folder, state, base model, size) or as covers. Chips by state, the same words as the scan page — Not scanned yet, Not found, Damaged, Found, Never scan — with counts; a search; the folders on the left with how many are picked in each, two folders of one type told apart by where they are. Pick these N picks everything shown (so "the 12 not scanned yet in LoRA" is two clicks), shift-click picks a run, Picked only shows the picks. The footer says how many are picked and how big they are.
  • Only files a scan can read (.safetensors) are listed; the old picker also offered .ckpt and other files a scan never opens.
  • Quick to open with many models, or models on a network drive (a NAS, a cloud server's model disk): the page reads the headers of models not scanned yet 16 at a time and remembers them until the file changes, so it opens again at once; while it counts, it says Counting your models… instead of staying blank. On a cloud server with 1041 unscanned models it had taken 79 seconds each time.

⚙️ Settings in tabs (设置分组)

  • The settings page is split into tabs: Look, Models, Workflows, Node packs, Prompts, Backup, Help. "Change where downloads go" opens Models directly.
  • Where downloads (and now imports) go is a choice: by base model (the new default), in Downloads then by base model, all in a Downloads folder, straight in the type's folder, or your own folder typed by hand; an example path shows the result. Without a saved setting, downloads now go by base model instead of into a Downloads folder.

🔖 Take single tags from a saved prompt (从存过的提示词里挑词)

  • In Prompt Studio a clicked card opens as a reference between the cards and the prompt you are editing, as its tags (with their meanings when a language is picked). Click one tag and only that tag goes into the prompt; tags the prompt has already are ticked. Add all puts the whole card in, as a click on a card used to; the preview's Add and dragging a card work as before. The ✕ closes the reference; beside the canvas the drawer widens by the column while it is open.
  • It does not squeeze a narrow window: where three columns do not fit (a side-panel browser window, a narrow drawer), the reference takes the card list's place and ‹ Cards brings the list back, so the reference and the prompt you edit each get half the width.

Fixes

  • The browser no longer opens blank after Put on canvas (搭配放到画布后再打开不再空白): putting a combo on the canvas hid the combo page before closing the browser, so the next time it opened the title said Combos and nothing was below. It now opens again on the combo list, or on the combo you put down.
  • Edit emotions: every clip in the list, and the line follows the clip (编辑情绪:下拉列出全部音频,换音频台词跟着换): a box that already held a file listed only the files whose names contain it (the browser filters by the box's text), so a character's other clips seemed missing. Pressing the box now lists all of them, with the current file shown greyed; leaving without a pick keeps it. Picking another clip no longer keeps the old clip's line beside it: the line the node already reads for that clip is filled in, otherwise the line the node finds for the clip (its .txt / .lab or the annotation list) is filled in, with Anomalous_TTS from interface 14 on (older versions leave the box blank and the node reads the line when it speaks); going back to the first clip brings its line back. A blank line box shows, greyed, the line the node uses for its clip.
  • Search a character's clips by line (按台词找参考音频): Edit emotions has a search box above the clips: type part of a line or a file name and the matches list with their lines, to hear (▶), Use as main voice or Add as emotion (a new emotion row with that clip and its line, ready for its name). Searching lines needs Anomalous_TTS interface 14; with an older version it searches file names.