Skip to content

AI media harness for planning, generating, and editing locally

TypeScript

Update Single Shot context guidance

Bittie2aa2beef5e3
diff --git a/RELEASE_NOTES.md b/RELEASE_NOTES.md
index 25b12bb..cc2bcef 100644
--- a/RELEASE_NOTES.md
+++ b/RELEASE_NOTES.md
@@ -9,7 +9,7 @@ Changes since 0.3.2.
- **Album covers:** Generate artwork from a song’s context menu using a dedicated generator selected under Settings → Generators → Icons.
- **Video library:** Organize Videos → Scenes → Shots, play scenes together, split songs into scenes, and share prompts and timed references. Videos can override project Look and resolution settings, and dependent scenes can be queued together. Cancelling a generation also cancels its downstream dependents while preserving earlier generations.
- **Scene takes:** Keep up to four takes plus a draft, restore settings, choose the primary take, and upscale saved takes by 1.5×, 2× or 4×. Non-primary takes can be removed during generation.
-- **Single Shot:** Continue with Context and Previous video anchors, optional generated endings, improved overlap joining and consistent Tone. Use Bridge for new connections instead of Long Shot.
+- **Single Shot:** Continue with Context and Previous video anchors, optional generated endings, improved overlap joining and consistent Tone. Context guidance now matches the original anchor to reduce appearance degradation across continuations. Use Bridge for new connections instead of Long Shot.
- **Motions and agents:** Create editable JavaScript animations for the timeline or references. Agents can inspect media frames, generate music and sync Motions to audio.
- **Image editing:** Combine or subtract lasso selections, queue batches of up to 100 images, and edit bounded regions of larger images while preserving their original resolution.
- **Live previews:** Watch MiniMax H3 video and image generation in the editor. Optional TAEH3 preview weights are available in Settings.
diff --git a/docs/slopfab-compatibility.md b/docs/slopfab-compatibility.md
index e7681ef..f01ac42 100644
--- a/docs/slopfab-compatibility.md
+++ b/docs/slopfab-compatibility.md
@@ -2,7 +2,7 @@
## Bundled runtime (API 1.48.0)
-Windows and Linux runtimes use SlopFab commit `fd264f6bc293d21e76cf26aaae3e70ed25f1e491`
+Windows and Linux runtimes use SlopFab commit `c5297c243923f336d4f174980a80452f0003bd0f`
from `D:\Projects\slopfab`, with CUDA and Vulkan enabled and FFmpeg disabled.
This update keeps Keystone and delayed Capstone context inside sampling: the
original front latent is restored before output, and Capstone adds one clean
@@ -15,7 +15,8 @@ Video previews autoplay and loop inside the existing shot card player; image
previews update the image canvas and thumbnail. Both use the generation previews
setting and require the optional TAEH3 weights. Saved-latent upscaling remains supported.
-LAN protocol is 28 so workers use the same image crop merging, Single Shot Tone and overlap behavior. Single Shot automatically shortens its preferred four-update
+LAN protocol is 29 so workers use the same Keystone conditioning, image crop
+merging, Single Shot Tone and overlap behavior. Single Shot automatically shortens its preferred four-update
Capstone delay when the resolved video schedule is shorter, without changing the
LoRA recipe. Restart Slopus and update workers to load the new runtime.
@@ -26,6 +27,23 @@ workaround documented below.
## Single Shot
+The bundled runtime includes the Keystone conditioning correction from SlopFab
+`89e1cca`. With Context video selected, the original context's first latent slice
+now guides both the locked target overlap and its visual conditioning at the
+same position. Previously the guide retained the advancing video's slice,
+contradicting the original anchor and weakening its influence early in sampling.
+The correction applies to CUDA and Vulkan and is intended to reduce appearance
+degradation across continuations. It requires no new C API calls. Context-video
+generation snapshots record the corrected guide so older takes can be regenerated;
+existing saved media and latents are not modified.
+
+Validation for this update includes SlopFab's host continuation tests, C API and
+embedded-tokenizer checks on Windows and Linux, and Slopus core and frontend
+tests. The bundled Windows DLL also passes the real-weight Keystone smoke test
+on CUDA and Vulkan, using Slopus's portable arithmetic policy for Vulkan. These
+small synthetic-prompt runs verify early anchor influence, source preservation
+and retained/file parity; they are not a visual quality benchmark.
+
With Context video selected, Single Shot continues from Previous video's saved
latents while using the sequence's first video as its persistent Keystone and
Capstone source. It runs one generation pass. The last 39 frames of Previous
diff --git a/lib/slopfab/libslopfab.so b/lib/slopfab/libslopfab.so
index 0a880ea..bbd8e91 100644
--- a/lib/slopfab/libslopfab.so
+++ b/lib/slopfab/libslopfab.so
@@ -1,3 +1,3 @@
version https://git-lfs.github.com/spec/v1
-oid sha256:5b62582deda0394b0c95f1a189646a7ad1363c8e679741562dae7ea34e81d4f2
-size 24045264
+oid sha256:9ff56c773b3786d6f7fb7f25853b58e354d1fdb6becb933c7cfccc948a2fc58a
+size 24086384
diff --git a/lib/slopfab/slopfab.dll b/lib/slopfab/slopfab.dll
index 639e671..dc418fb 100644
--- a/lib/slopfab/slopfab.dll
+++ b/lib/slopfab/slopfab.dll
@@ -1,3 +1,3 @@
version https://git-lfs.github.com/spec/v1
-oid sha256:b7a21c7ac4c14e6ab713711ee7aee332c5fabb151d845551f410976b89713649
-size 20421120
+oid sha256:180652874381eee566c51da8e77ce3444dcd6ed175b3f2a4c15532d7c8cb6cd1
+size 20452864
diff --git a/src-tauri/core/src/worker/protocol.rs b/src-tauri/core/src/worker/protocol.rs
index 8d01221..edf69ca 100644
--- a/src-tauri/core/src/worker/protocol.rs
+++ b/src-tauri/core/src/worker/protocol.rs
@@ -32,7 +32,8 @@ pub const SERVICE_TYPE: &str = "_slopus-worker._tcp.local.";
// v26 replaces Single Shot's 39-frame overlap, holding 22 and regenerating 17.
// v27 grades Tone once after joining Single Shot's neutral ending and main clip.
// v28 bounds image-edit inference to the target region and preserves the original canvas.
-pub const PROTOCOL_VERSION: u32 = 28;
+// v29 aligns Single Shot's Keystone conditioning guide with its original context anchor.
+pub const PROTOCOL_VERSION: u32 = 29;
/// Streaming SeedVR2 restoration; model paths are resolved on the worker.
#[derive(Debug, Clone, Serialize, Deserialize)]
@@ -387,7 +388,7 @@ mod tests {
let single = restored.single_shot.unwrap();
assert_eq!(single.ending_prompt, "The subject waves goodbye");
assert!(single.ending_path.is_none());
- assert_eq!(PROTOCOL_VERSION, 28);
+ assert_eq!(PROTOCOL_VERSION, 29);
}
#[test]
diff --git a/src/lib/project.singleShot.test.ts b/src/lib/project.singleShot.test.ts
index 899156b..433c796 100644
--- a/src/lib/project.singleShot.test.ts
+++ b/src/lib/project.singleShot.test.ts
@@ -17,7 +17,7 @@ it("keeps the original context archive while continuation advances and persists
const request = sceneGenerationRequest(next, config, "C:/Project", 9);
expect(request).toMatchObject({ contextSceneId: "original", keystoneRelativePath: original.latentRelativePath,
previousSceneId: "previous", continuationRelativePath: previous.latentRelativePath, singleShot: { endingPrompt: "" } });
- expect(JSON.parse(sceneGenerationSnapshot(next, request))).toMatchObject({ contextSceneId: "original", capstoneLockAfter: "auto" });
+ expect(JSON.parse(sceneGenerationSnapshot(next, request))).toMatchObject({ contextSceneId: "original", keystoneGuide: "context", capstoneLockAfter: "auto" });
expect(sceneGenerationSnapshot({ ...next, endingPrompt: "Something else" }, request)).toBe(sceneGenerationSnapshot(next, request));
expect(parseProjectConfig(JSON.parse(JSON.stringify(config))).generationJobs[0].contextSceneId).toBe("original");
expect(sceneInput(next).contextSceneId).toBe("original");
diff --git a/src/lib/project.ts b/src/lib/project.ts
index 8a4ecde..c7eaf09 100644
--- a/src/lib/project.ts
+++ b/src/lib/project.ts
@@ -1307,7 +1307,7 @@ export function sceneGenerationSnapshot(job: GenerationJob, input: SceneGenerati
...(job.sceneType ? { sceneType: job.sceneType } : {}),
prompt: input.prompt,
...(input.loraSnapshot?.length ? { loras: input.loraSnapshot } : {}),
- ...(input.contextSceneId ? { contextSceneId: input.contextSceneId, keystoneRelativePath: input.keystoneRelativePath, capstoneLockAfter: "auto" } : {}),
+ ...(input.contextSceneId ? { contextSceneId: input.contextSceneId, keystoneRelativePath: input.keystoneRelativePath, keystoneGuide: "context", capstoneLockAfter: "auto" } : {}),
...(input.singleShot ? { singleShot: input.singleShot, ...(!input.contextSceneId ? { endingPrompt: job.endingPrompt ?? "" } : {}) } : {}),
...(input.videoTransition ? { videoTransition: input.videoTransition } : {}),
...(input.latentBridge ? { latentBridge: input.latentBridge } : {}),