- api/ai_gates.py: changing the canvas, using the GPU, downloading and scanning online can each
go ahead, ask first or never (downloads ask by default; anomalous/ai_gates.json). A call that
asks waits on its step (state "asking", what it would do listed) up to 50 s for Allow or No
(with the user's reason for the AI); unanswered, the AI is told to wait and call again, the
later answer reaches it as a message, and the same call after an Allow goes ahead.
- ui_ai_gate.js: the card under a waiting step and the Gates window; a toast while elsewhere.
- ui_ai_flow.js + ai_task_view.flowStages: the task's stages in order, each done, running,
waiting or failed; a stage leads to its first step.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- api/action_catalog.py: every MCP tool's kind, stage of the work and gate in one place (a test
keeps it and the tools in step); steps now carry their stage.
- api/ai_task_changes.py: changes the user makes in Anomalous while an AI works become steps of
its task, reach the AI once with its next call (kind "change": keep it), and the activity
entry names the task; the activity page links to it ("AI task ->").
- api/ai_task_runs.py: run watching moved out of ai_tasks.py (600-line rule); a step a restart
left "running" is marked stopped when the record is read again.
- Pictures: inspector / compare with canvas / keep as tools on hover instead of three buttons
under every picture; tick any two (any step or task) and compare from the bar at the foot of
the page; the results strip shows thumbnails again.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Panels are shown on the stage beside the timeline (ui_ai_side.js), so they
stay in view however many steps follow; the timeline keeps one line per
panel that brings it there. The stage follows the newest panel until the
user goes back to another (pager or a panel's line). Too narrow for three
columns, the timeline and the stage become two tabs; a panel that asks
brings the stage forward. The task list can be folded away. What a step
made moves to ui_ai_results.js; the toast goes once the workbench opens.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
New panel blocks, data only and drawn by Anomalous: marks on pictures
(circles or boxes at fractions of the picture, numbered, labels under it
that zoom there), a before / after slider, picks (pictures with the AI's
reason, Take this one as an answer, Put on the canvas through ComfyUI's
own file loading, drag onto the canvas) and diagrams (boxes and arrows laid
out by ai_stage_layout.js; a box naming a canvas node finds it). No SVG
or drawing code of the AI's is taken.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
show_panel lays out a panel as a step of the AI's task from fixed blocks
(heading, text, pictures in a row, grid or side by side, the user's models
as cards with covers, table, list, params, progress, divider) and can ask:
buttons (one or several) and a line of text. api/ai_panels.py checks every
block (only these kinds, bounded plain text, pictures only from the output
folder, models only from the user's folders); nothing is run or read as
markup, and ui_ai_panel.js writes text only. Answers come in through
POST /anomalous/ai_tasks/{id}/answer and wait for get_answers (which can
wait up to 55 s); update_panel changes or closes a panel. A toast says
when a panel comes while the browser is elsewhere.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A new rail page keeps each MCP call as a step of a task (api/ai_tasks.py,
one file per task in the user folder, the newest 300): named by the AI
with the new begin_task tool, else one app's calls without a long pause.
A step keeps what it did and what it made (models, the combo, the canvas
change by its activity entry, the pictures a run made, watched in the
history until it ends, spoken audio, scans); the new report tool leaves
the AI's notes and pictures it made another way. The page shows the tasks
and one task's results strip and timeline with links (picture viewer,
model page, combo, find or undo on the canvas, audio), updates live, and
gives the rail a dot while you are elsewhere; Home lists the last tasks.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Put on canvas hid the combo page before closing the browser, so the next
open showed the Combos title over nothing; now only the browser closes and
it opens again on the list or the combo. A character with no line in the
script hears a sample sentence on ▶, which the panel now says; on the
whole script ▶ reads the line with its own character.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
🗑 on a row asks first (saying where the clip is used), then has the node
move it with its own .txt / .lab to the Recycle Bin (never a permanent
delete) and take it out of the settings; the window drops it from its
voices too. Needs Anomalous_TTS interface 18.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
"Reference audio and emotions" replaces Edit emotions: a large window with
the voices on top (main voice, emotions, new emotion) and every clip of the
character below, one row each (listen, file and length, its line editable
in place, where else it is used, Main / Aux for the voice picked), with
filters, search and paging. Auxiliary references (timbre averaged in, v1 /
v2 / v2Pro), fixed lines (`lines`) and lengths need Anomalous_TTS
interface 18. tts_voice_refs.js holds the rules; the old clip search and
merge helper go.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Every script is a draft, saved on its own in ComfyUI's user folder
(anomalous_model_browser/voice_over_drafts.json through /userdata), read
again and joined before each write so two tabs keep each other's drafts.
A bar over the script names it and leads to New, Drafts and Close; the box
shows a card per draft (open, rename, copy, delete) and Recently deleted
(30 days). A reload opens the draft that was open; a file voiced again
from the gallery becomes a draft of its own; the one localStorage draft
of the earlier build moves in once.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The side panel changes one layer at a time under a path (Whole script ›
character › line); each layer has the same rows (speed, tone, language),
each saying where its value comes from, with ↺ to follow the layer above.
The more particular wins: line, character, whole script, the character's
usual settings. The whole script's values go into the node's inputs and
only lines that differ get tags. Lasting settings are links at the foot.
The script is kept as a draft in this browser and comes back after a
reload; editing the text keeps unchanged lines' own settings.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
GPT-SoVITS Settings gets a status chip by the title and tabs (tones, other
languages, base files with a dot when they need the user, where things are,
version); the tones have a Simple / Numbers switch. Voice-over's side panel
gets How it reads / Numbers / Settings tabs for a character or picked lines.
The More fold is gone; ui_tab_strip.js replaces ui_tts_fold.js.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Voice-over's side panel shows who reads, speed, the language (one line
while detected) and the tone with a listen button; the comparison, the
numbers and the links fold under More. GPT-SoVITS Settings opens with one
status line and the tone to use; the language rules, base files, places
and version are folded and open by themselves when they need the user.
Plainer words throughout: tone, Steady / Natural / Lively, As the
character, Always read like this, Own settings, base files.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The side panel follows what is picked: the whole script, one character in
it (its language, speed and mode for this script), or the picked lines (who
reads them, and their own values, each following the character until set).
A line set for itself carries a tag on its card. The values reach
Anomalous_TTS interface 17 as [lang:…][speed:…][preset:…] on each line;
an older node reads the whole script as before. Clicking a character in the
cast now shows it instead of giving it the picked lines.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
69 rules and 21 selectors in lists named plugin classes no script makes (old paper-note
cards, the old drawer header, the combo editor's old toolbar, icon buttons, tag colours…).
No behaviour or look changes.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The combo editor gets a head with its picture (See the picture), a large name with a
Rename button, chips of what it holds and Put on canvas; models show only what is chosen,
Change / Add open the grid of compatible ones. No Save button: everything saves itself
(a tag changed in the bilingual view now saves too). The gallery's ★ Kept lists only the
pictures kept as a workflow, combo or prompt.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A bar over the canvas finds combos, models and saved prompts as you type
(GET /anomalous/quick_index lists the models in a few words) and puts the chosen one
where the pointer was. Also in the empty canvas's right-click menu, now registered.
Saved model paths in a linked model folder get their real place.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A node's menu offers Open in AMB, change model by picture, insert a LoRA before / after,
saved values, the model's pictures and details, save values and save as combo, each where
it applies; the empty canvas's menu opens AMB's pages. Entries can be switched off, moved,
and put in one AMB submenu or right in the menu. The model picker's Esc closes only the
picker, wherever the focus is.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The card shows the output image with the main model's cover as a badge; the name is the
first LoRA (else the main model) and two telling tags; the toast after keeping shows the
picture and renames the combo on the spot, and the editor's title renames on a click and
opens the picture.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Every line is read by its own character: pasted and added lines go to
the one picked in the list on the left, the node's character is the
first line's, and the "Who speaks" bar is gone. Cards (ui_script_cards.js)
show the text as text: a click picks, a double-click or the pencil edits
it in place (Enter keeps, Esc puts back), and the name on a card opens a
menu of characters (all picked lines when it is one of them). Language
and speed are no longer offered to keep when several characters read;
the settings link names the listened line's character.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Pick lines on Voice-over (click; Ctrl / Cmd adds, Shift a run; Esc lets
go) and click a character in the list on the left to have it read them;
with none picked, the click changes who reads by default. On Voice-over
the list is the cast: narrower, characters with their colours, and a
click never leaves the page. Cards show who reads each line in its
colour, with that character's emotions; the bundle carries [name] where
the speaker changes, so generating, the cards' listen button, pushing,
dragging and copying follow it, and a file voiced again from the audio
gallery keeps its speakers. The character cards' Voice-over button is
gone. The workflow actions move to ui_script_workflow.js and the cast
helpers to ui_script_cast.js, keeping the director under 600 lines.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Each card has a ▶ (on hover, and on the card you are on) that reads its
line in the mode picked on the side; the side names the line and keeps
the modes heard on it as buttons to compare, without a listen button of
its own. The cards are drawn again once the node's modes are known.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A full-window button beside Dock (and a double-click on the title bar)
makes the window fill the ComfyUI page, remembered; it is neither dragged
nor resized then, and docking turns it off. Voice-over: one row per line
card, a divider between the lines and the generating side (dragged and
remembered, ui_split_bar.js; about half each by default), "Into your own
workflow" folded, who reads and their voices on one row, the how-to line
only until there are lines. On a wide side the modes' scale and numbers
are shown open unless the fold was closed before.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The line card you click or edit becomes the one the modes are heard on:
it is marked "heard on the right", and the run section names it beside
Listen ("line 3: ..."), replacing the dropdown. An empty line falls back
to a sample in the character's language.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
On the narrow side the scale joins the numbers in one fold and the long
notes became one sentence with the mode's description, so the side fits
without scrolling at a mid-size window; the side is a little wider. A
dropdown beside Listen shows which line of the script is read (#1 at
first) and picks another; the takes heard follow that line.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A bordered input pulled the eye from "15 -> 10". The number now looks like
plain bold text with a dashed underline; the box shows on hover or while
typing, without spin arrows (they covered the digits). The old value is a
little brighter so the change reads as one line.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A mode is shown against one fixed mode (the default preset in Settings,
Auto on Voice-over): a steady-to-lively scale with both marked and a
sentence ("a little livelier than your default"); the numbers move into a
fold. Listening reads one line with one seed through Preview Audio, only
the mode changing; takes are kept, and the modes heard stay as buttons to
compare. Settings picks the character and the line; Voice-over uses the
script's first sentence.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
GPT-SoVITS Settings and the Script Director share one switch: the modes
as choices, the picked one's numbers drawn like the Current node card
(two bars, before -> now, an arrow), each with its plain and its node
name, editable in place and saved into the preset. Settings gets Make
default / Reset / + New / Rename / Delete around it. The Script Director
replaces its folded fine-tuning numbers (which showed the node's defaults,
not what the presets used) with Auto or one preset for the whole script,
sent as the node's preset input, plus links to the presets and to the
character's generation settings; Save as default keeps the language and
speed and leaves the rest of the character's defaults alone.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The numbers table marks values that differ between presets and changed
built-in values (with the built-in one), like the compare window; built-in
presets reset one by one or together. New presets start as a copy, are
named by the user and can be picked everywhere. A character's generation
settings show each row's numbers in effect next to the general ones.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
For Anomalous_TTS interface 16: the default preset (Steady / Standard /
Lively), a preset for each pair of languages the characters can meet and the
presets' numbers, saved at once to the node's presets file; a Generation
settings button on character cards for the character's own preset, speed and
numbers, as a whole and per language. Older nodes get a pointer to the update
line instead.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The version manager takes an optional Checkout, so the companion node pack is
handled like the plugin itself: api/tts_version.py finds it among the
custom_nodes folders by its node class and accepts only a git clone of the
official repository; releases are fetched only on "Check for updates", an
update checks out the newest stable tag and can be undone. A checkout without
the release tags (cloned before it, then pulled) is matched by ls-remote.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Character cards show the version Anomalous_TTS reports (interface 15). The
v3/v4/v5 base models and vocoders are listed, counted and reminded of only
when a character or a dropped SoVITS file needs them; the import tells an
unreadable SoVITS file from a version an older node cannot use.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A combo saved from picked nodes shows each model as a card like the model
library's: the cover with its slot written on it, the name and node below,
a click on the card (or Change on the cover) picks another model. The
default combo's main models and LoRAs to pick from are cover cards side by
side instead of wide rows; its empty cover is a class, not inline styles.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
With many models, or models on a network drive, every visit to the scan
page read every folder and, after a restart, every new model's header
again (79 s for 1041 models on an AutoDL model disk).
api/model_heads.py keeps what each file's header says by real path, size
and time in user/anomalous/model_heads.json (at most 50,000 files, the
longest unused dropped first, written a few seconds after the last
change, only once setup_routes called keep_on_disk); get_metadata reads
headers through it, so a header is read once per file version.
scan_summary keeps the last summary (memory, and scan_summary.json when
it changed); ?quick=1 hands it out at once and the page shows it with
"showing the last counts while they are updated", then the fresh ones.
Requests meanwhile await one count (_fresh). The models found on Civitai
(components apart, at most 5000 kept) are listed by GET
/anomalous/scan_matched; the page's "Found on Civitai" count opens them,
with Civitai's name, base model, a box to find one by name and Open.
Checked on the isolated ComfyUI reading the real loras and checkpoints
read-only (scans stubbed; no file there changed): the last counts showed
at once also after a restart, the fresh ones replaced them, the found list
had 68 models with names, the box filtered, Open and Back returned to it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
On a cloud server with 1041 models not scanned yet, on the platform's
network model disk, GET /anomalous/scan_summary took 79 s, each time the
page opened, and the page stayed blank meanwhile: it read each new
model's header (for "looks like no drawing model") one after another.
The headers are now read 16 at a time (HEAD_READERS) and through
get_metadata, whose cache keeps each answer until the file changes, so
the page opened again reads none of them and the picker after it finds
them there. Opened afresh, a count slower than 400 ms shows "Counting
your models…" instead of a blank page.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The header's button rules are !important, so a <button> chip lost its own
colours and size; it is a span again, with role="button", a tab stop and
Enter / Space opening its explanation. The start-up line about
allow_remote_changes is flushed at once, so it shows even when ComfyUI's
output goes to a file.
Checked on an isolated ComfyUI behind a stand-in gateway that adds
X-Forwarded-For (as a cloud platform's proxy does): without the file the
page is view only and changes answer 403; with it the chip reads "remote
changes on" in yellow, a scan choice is saved and reset, and a picked
offline scan runs to the end.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
ComfyUI on a rented cloud server (AutoDL, RunPod) is only ever opened
from another device, so since v1.58.1 the plugin was view only there.
An empty allow_remote_changes file in the plugin folder, made on the
server and read once at start, now makes local_guard.may_change true for
every request; the middleware and the routes that check for themselves
(import, placement, scan choice, recipe focus) use it. is_local stays for
import's search of this computer's Downloads and for MCP. The page cannot
make or remove the file, and the file is git-ignored.
GET /anomalous/access answers {local, changes}. The header chip is now a
button: "view only" says what is refused and the two ways for cloud users
(SSH tunnel, the file); with the file it reads "remote changes on", in
yellow, and says that anyone with the address may change things.
docs/guides/cloud-server.md explains both ways and the risk; README,
AGENTS.md rule 4, backend.md and the CHANGELOG (v1.61.0) say so too.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The rule that hides the list column while a workspace is open matched
the workspace panel's style attribute inside :has(). Chrome then treats
every inline-style write on the page as a possible change and restyles
the whole page, so extensions that move their buttons every frame took
a large workflow's canvas down to 2-3 fps. The container now gets an
anomalous-workspace-open class from a MutationObserver on the panel's
style (every module opening a workspace sets its display), and the CSS
keys off that class.
Found, measured and fix proposed by Lindemellis in issue #22.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
New users did not know to scan first. Until anything is scanned, Home shows a card with four
steps and "Set up for me": GET /anomalous/civitai_check (api/network_check.py, local requests
only) says whether Civitai answers and whether a set key is refused; when it does not, the card
offers Retry (with a proxy) or an offline scan. The scan never renames files or refetches,
whatever the scan page was set to; the card follows its progress. Never a popup; "Later" hides it
per browser and Home's "New here?" line brings it back.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Stacking many LoRAs made a long chain of Load LoRA nodes. A default combo now says which LoRA
node it is laid out with (data.loraNode: ComfyUI's own chain, rgthree's Power Lora Loader, or
Easy-Use's LoRA stack + Apply LoraStack), chosen under New or in its editor; the third-party ones
are offered only when installed, and a missing one falls back to ComfyUI's own and says so.
Save as combo lost a Power Lora Loader's rows (object values were not stored): they are kept as
the node's rows, made again before the slots are filled, each row's LoRA a slot.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A prompt quoted on one line in a Civitai description made the whole details column scroll sideways.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
