Repository navigation
Scumble 0.1.53: each area on its own, Krea 3 Edit, and SAM2 off the main thread
Get Scumble
| Package | Download | Size | Notes |
|---|---|---|---|
| Microsoft Store (Windows) | Scumble in the Store or winget install scumble |
– | Signed by Microsoft and updated by the Store; a new version can reach the Store a few days after GitHub (Microsoft's certification). |
| Windows installer | Scumble-Setup-0.1.53.exe | 124 MB | Updates itself; not code-signed yet, so SmartScreen asks once. |
| Windows portable | Scumble-0.1.53-portable-win-x64.zip | 175 MB | No install: unpack and start Scumble.exe; keeps its data in a folder beside it and does not update itself. |
| Linux AppImage | scumble-0.1.53.AppImage | 152 MB | Early build made by CI, not tested by the author yet. |
| Linux .deb | scumble-0.1.53.deb | 120 MB | Early build made by CI, not tested by the author yet. |
| Claude Desktop extension | scumble-0.1.53.mcpb | 81 KB | Adds Scumble's MCP tools to Claude Desktop with one click; needs Scumble installed (any of the above). |
The other files of this release (latest.yml, latest-linux.yml and the .blockmap) are for the app's updater; you do not need to download them.
- Krea 3 Edit, on fal.ai, WaveSpeedAI, Replicate, Comfy Cloud, OpenRouter, Comfy Router and Oxen.ai, with up to eight pictures, a Planning setting and Generate new (not on OpenRouter). It repaints the whole crop and your selection limits what lands. Built from the providers' docs, not run against the live APIs yet.
- Cancel on fal.ai and Replicate now asks the provider to stop the job. Before, Scumble only stopped waiting and the job ran on and was charged; a run that times out asks the same. If the provider does not answer within two seconds the line says the job may still finish.
- Importing a recipe no longer overwrites another one without saying so. A workflow or a recipe without an id of its own gets a free name, and a recipe that does carry an id says whether it replaced one of yours or shadows a shipped one, in the status line when you import from the File menu too.
- Long API errors stay readable. An error answer no longer carries a whole picture back into the status line and the log, and a key inside one is removed before the text is cut.
- A fill layer with a depth limit exports almost twice as fast, and with edge snap its soft edge now follows the picture as it does on screen. On a 15,000 x 10,000 picture the fill layer takes about 6.2 s instead of 11.2, a PSD holding one 9.9 s instead of 16.7. As long as the picture below is still the one the depth map was made from, export and screen agree exactly up to 4096 px on the long side; on larger pictures a hard colour edge can differ in a band a few pixels wide.
- A crash or a locked file can no longer wipe your autosave, settings or API keys. The three stores are now flushed to disk before they replace the old file, and a read that fails no longer counts as an empty store. A damaged
settings.jsonorsecrets.jsonis kept under a name of its own (settings.json.damaged, then.damaged.1and so on), the status line and the log say so once, and new settings and keys save normally again. Closing the window while another program holdssettings.jsonno longer skips the save. - Each area of a selection on its own (issue #9): a selection of several separate areas (three people selected by text, say) goes to the model as one crop around all of them, scaled down to the model's size, so each area came back soft. Before such a run the status line now says so when that crop goes out at less than half its size, and Each area, beside Generate, is shown while the selection holds two or more separate areas: it generates each area as its own crop and run with the same prompt and recipe, one after another, each result its own layer, and leaves your selection as it is (a turn, a crop or resize, Generate new and a second Generate on that tab wait until the areas are done). Up to 8 areas; areas whose crops would overlap count as one, and specks (smaller than a 20,000th of the picture, and anything under 256 pixels) outside the areas are left out. On an API model each area is a run of its own and is charged as one, so Scumble asks first and names the runs (the in-app models are free and do not ask). On your own ComfyUI the next area is queued only once the one before has landed. While the areas run the button reads Stop: it stops the run going (an API run as Cancel does; one on your ComfyUI finishes there) and the areas not yet sent. For agents:
generatetakeseach_area(its timeout bounds the whole run),statuscounts the selection's areas, and the in-app assistant's question names the runs. The areas are counted in the background on a coarse copy of the selection: on a 15,000 x 10,000 picture the window spends under a millisecond on it. - The blur and sharpen brushes and frequency separation take about half the time, with the same pixels as before. On a 15,000 x 10,000 picture a 400 px blur stroke takes about 33 ms per move instead of 74, a 1000 px one about 99 instead of 231, the sharpen brush about half as long as well, and frequency separation of a 16-megapixel area about 0.7 s instead of 2 s, while the window stops answering for half as long. The soft edge of a generated result's mask and relight's surface map are worked out a little faster too.
- The pixel kernels really give memory back now. In 0.1.52 the first kernel instance of the window and of each background worker stayed in memory after it was replaced, so the memory 0.1.52 said its workers returned after a large operation (about 290 MB after a 15,000 x 10,000 export) stayed in use until the app was closed; now it is freed.
- GPT Image 2.6 Sunburst and Flare on Comfy Cloud too, through OpenAI's partner node, with a Complexity setting and the selection as the mask (built from the node's own description, not run on Comfy Cloud yet).
- GPT Image 2 and 2.5 on Comfy Cloud: a run with a selection and reference layers or the Original no longer fails. It goes without the mask, which that node takes beside one picture only, and the selection still limits the result; the status line says so as a plain note.
- A filter layer's sizes follow a resize. Resize image, an upscale of the whole picture and the Realism Pass above 1x now scale a filter layer's sizes in pixels (a blur's radius, a Grain layer's size and plate scale, the glass and chromatic shift, halation, glow, tonal contrast and structure radii) by the same factor, kept within each slider's range, in the same undo step, so the effect keeps its strength on the new size. The grain a Film look or a black-and-white film adds of its own keeps its size in pixels, as do layers of a plugin that is off (the status line says how many).
- Seedream 5.5 on BytePlus ModelArk (early access: BytePlus has to let your account in, then you activate it in the ModelArk console under Model activation). It edits with up to ten pictures and makes new pictures with Generate new, at 0.92 to 4.6 megapixels, as PNG. Its Background row set to Transparent asks for a cut-out, and the result layer keeps that transparency inside the selection. Written from BytePlus's preview docs, not run live yet.
- The dodge / burn and sponge brushes are faster, with the same pixels. On a 15,000 x 10,000 picture a 1000 px sponge stroke moves in 67 ms instead of 87 (a 400 px one in 22 instead of 31), and a 1000 px dodge in 64 instead of 69.
- Liquify lands faster, with the same pixels. The bake of a stroke runs in Rust: on a 15,000 x 10,000 picture the pool's work for a 1000 px push is about 1.8 times faster, and a 200 px push at 1:1 lands in about 20 ms instead of 25. The live preview while you drag runs in Rust too: a 1000 px push at 1:1 draws a frame in 10 ms instead of 16, and the brush itself works about twice as fast (a 1000 px push on a fitted 15,000 x 10,000 picture: about 5 ms a move instead of about 14).
- A local run can no longer pick up another picture of the same name on your ComfyUI. Pictures you open, paste or drop now go up under a name made from their content, such as
image~3fa9c1d2.png, so two different pictures namedimage.png(from a second copy of Scumble, the Inpaint Canvas node or an older session) no longer get mixed up. Layer and tab names stay as they were. A document saved earlier still works: if the server holds a different file under its old name, the run uses its own copy. - A filter layer no longer thickens soft edges. A cut-out or other see-through picture under a filter layer looked more solid at its edges on screen and in some exports than in the plain PNG. They now agree. Opaque pictures look the same, apart from a very small shift (a level or two) inside the soft ramp of a filter layer's Limit.
- A paid run on fal, Replicate or WaveSpeed no longer ends on a short connection problem. Scumble now rides out a few failed status reads and retries a failed download of the finished picture, on Comfy Router and the Partner API too. If it does give up, the message names the job's request id, so you can find the picture at the provider.
- Invert selection is about twice as fast. On a 15,000 x 10,000 picture it takes 0.13 s instead of 0.23, with the same selection. The magic wand over a whole 15,000 x 10,000 picture lands in 0.83 s instead of 1.0, and shrinking a selection by 16 px in 0.35 s instead of 0.48.
- A large ORA exports faster. A 15,000 x 10,000 picture as ORA takes about 30 % less time: the checksums of its entries now come from the pieces the writer has already made, instead of a second read.
- Select by depth is two to three times as fast on large pictures, and select by brightness or colour got much faster too. On a 15,000 x 10,000 picture a selection by depth lands in about 0.2 s instead of 0.60, with edge snap in about 1.1 s instead of 2.3, by brightness in 1.2 s instead of 1.6 and by colour in 1.8 s instead of 3.1: areas that are wholly in or out of the range are no longer worked through pixel by pixel. The selected pixels stay the same, with two exceptions: with edge snap on a picture with see-through parts, the edge no longer picks up leftovers of the row above, so it can differ there by a few levels; and a pixel lying on a rounding boundary can move by one level of 255.
- A TIFF with damaged LZW data after the picture's own bytes now opens instead of sometimes being refused.
select_range,add_filterandset_filterrefuse a malformed limit (a colour that is not#rgbor#rrggbb, a bound that is not a number) with a readable message. A document that already holds such a limit keeps drawing as before.- Generated results land a little faster, with the same pixels. The masks a result is stitched in with are built in Rust: a provider run's crop of a 15,000 x 10,000 picture is ready in 0.62 s instead of 0.72. The colour match that fits a result to the picture around it runs in Rust too: a 4096 px result is stitched in 2.3 s instead of 2.7. Fitting a result's edges to the picture got faster too: a 2048 px result with its edges aligned is ready in 0.61 s instead of 0.75.
- Panning a large picture at 1:1 is smoother. The upload of tiles to the GPU copies opaque and empty areas whole: a far pan at 1:1 over a 15,000 x 10,000 picture takes 6.9 ms a frame instead of 12.1, and the first frame after a whole change at 1:1 21.7 ms instead of 29.2, with the same pixels.
- Settings opens at once. The GPU memory line fills in a moment later instead of holding the dialog back (about 70 ms on a machine with an NVIDIA card, seconds on one without its tool).
- Keying: refining a green or blue screen key larger than the model's size fits its matte about 18 % faster, with the same matte.
- The curves panel's histogram is the luma of the picture below the layer, on both pixel engines and with the GPU filter chain; transparent areas no longer count as black, and filter passes no longer read a thumbnail back for it on every redraw.
- Clean up no longer removes what an undo can bring back. Settings › Local files › Clean up kept only the files the open documents showed, so a deleted layer, a flattened or cropped picture, a named snapshot's layers, a user font or what the assistant's Undo this turn holds could be removed while Ctrl+Z still brought them back, and the next start could no longer find that layer or the whole picture. Now every file an undo, a redo, a snapshot or the assistant's turn still holds is kept; a file that is gone anyway is stored again when its layer or picture comes back; a layer whose file is missing when a document loads is named in the status line, and a text layer in that case is drawn again from its text. A closed tab no longer stays in memory after a menu, a choice list or a question was opened in it.
- Selecting objects no longer freezes Scumble. SAM2, the background-removal models and depth now run in a process of their own, so the window, agents over MCP and the autosave no longer stall while they load or run (the first SAM2 call held everything up for about a second, a cutout on the CPU for about five); the results are the same as before. Free VRAM gives their memory back once a job under way has finished, and after five minutes without a helper call the process ends by itself. If that process stops twice within a minute or cannot start, the status line says so once: the helpers then run in Scumble's main process for the rest of the session (slower; restart Scumble to try again).
- Pictures with fine noise in every channel are saved whole again. Since 0.1.17 a picture with near-uniform noise in all four channels, transparency included (an imported RGBA noise texture, say), about 700 px wide or more, could be saved cut short: every 256-row piece of its PNG but the last lost its final bytes, so such a layer in the local store (and with it the .scumble and the autosave), a flattened picture and PNG and ORA exports reopened with only their top rows. Photos, renders, generated results and cut-outs were not affected. Files written before keep the damage: an affected layer opens showing only its top part.
- Hardening. Links leave for the browser only as http(s); Scumble's window navigates only to its own page; the ComfyUI window's page gets no location, MIDI or device access and cannot hand other schemes to the system; when it asks for your camera or microphone (ComfyUI's Webcam Capture and Record Audio nodes do), Scumble asks you first, Allow or Deny, once for each until it quits, only for the page of the ComfyUI you set or Comfy Cloud's own page, never for anything it embeds (the question takes no click for its first half second, and Deny is its default); a crafted PSD or ORA (a deflate bomb, a huge layer or mask, overlapping channel data, layers past the picture's size) is refused with a plain message instead of filling memory; the local command socket reads long lines in linear time and caps them at 500 MB; recipe ids are file names (made safe when read, refused at save; Save and Remove use the recipe's own file); Settings › ComfyUI's Test asks its questions together, so it answers sooner; the RunPod template's node packs are pinned and its Hugging Face token goes by curl.
Scumble --run: batches from a terminal.Scumble --run <steps.json>(or--run -for stdin) runs a list of commands in one call on one Scumble: the open one when it runs, else one it starts without a window. It can run the steps once per file of a folder, with{file},{stem},{ext}and other placeholders, prints a JSON line per step and exits 0 only when nothing failed;--keep-goinggoes on with the next file after a failure. A document you already have open is left alone, and the batch's own tabs stay off Reopen Closed Tab. Three examples are indocs/examples(cut out, key and a film look on a folder). For agents,close_documenttakesremember(default true; false keeps the tab off Reopen Closed Tab).- Snapping to other layers. A layer you move or scale with the transform tool now also snaps its edges and centre to the edges and centres of the other visible layers, within the same 8 screen pixels and with the same dashed line; a layer still being turned snaps by the box you see, and hold Alt to place it freely.
- The eyedropper can average. A Size row in its options (Point, 3 x 3, 5 x 5; Point by default, remembered) picks the mean of the pixels around the cursor, so grain or noise does not give you one stray pixel's colour. Alt+click with the paint brush follows it; at the picture's edge only the pixels inside count, and transparent pixels count by their opacity.