@koda-sl/baker-cli 0.298.0-dev.52a1f12f8 → 0.300.0-dev.e0610b71d
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
|
@@ -1211,6 +1211,39 @@ var GEMINI_OMNI_IMAGE_MIMES = ["image/png", "image/jpeg", "image/webp"];
|
|
|
1211
1211
|
var GROK_IMAGINE_DURATIONS = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10];
|
|
1212
1212
|
var WAN_27_DURATIONS = [2, 3, 4, 5, 6, 7, 8, 9, 10];
|
|
1213
1213
|
var HAPPYHORSE_DURATIONS = [3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15];
|
|
1214
|
+
var WAN_30_DURATIONS = [
|
|
1215
|
+
2,
|
|
1216
|
+
3,
|
|
1217
|
+
4,
|
|
1218
|
+
5,
|
|
1219
|
+
6,
|
|
1220
|
+
7,
|
|
1221
|
+
8,
|
|
1222
|
+
9,
|
|
1223
|
+
10,
|
|
1224
|
+
11,
|
|
1225
|
+
12,
|
|
1226
|
+
13,
|
|
1227
|
+
14,
|
|
1228
|
+
15,
|
|
1229
|
+
16,
|
|
1230
|
+
17,
|
|
1231
|
+
18,
|
|
1232
|
+
19,
|
|
1233
|
+
20,
|
|
1234
|
+
21,
|
|
1235
|
+
22,
|
|
1236
|
+
23,
|
|
1237
|
+
24,
|
|
1238
|
+
25,
|
|
1239
|
+
26,
|
|
1240
|
+
27,
|
|
1241
|
+
28,
|
|
1242
|
+
29,
|
|
1243
|
+
30
|
|
1244
|
+
];
|
|
1245
|
+
var GROK_IMAGINE_15_DURATIONS = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15];
|
|
1246
|
+
var RUNWAY_45_DURATIONS = [2, 3, 4, 5, 6, 7, 8, 9, 10];
|
|
1214
1247
|
var ELEVENLABS_OUTPUT_FORMATS = [
|
|
1215
1248
|
"mp3_22050_32",
|
|
1216
1249
|
"mp3_44100_32",
|
|
@@ -1248,7 +1281,10 @@ var VIDEO_GENERATE_MODELS = [
|
|
|
1248
1281
|
"google/veo-3.1-lite",
|
|
1249
1282
|
"kwaivgi/kling-v3.0-pro",
|
|
1250
1283
|
"x-ai/grok-imagine-video",
|
|
1284
|
+
"x-ai/grok-imagine-video-1.5",
|
|
1251
1285
|
"alibaba/wan-2.7",
|
|
1286
|
+
"alibaba/wan-3.0",
|
|
1287
|
+
"runway/gen-4.5",
|
|
1252
1288
|
"alibaba/happyhorse-1.1"
|
|
1253
1289
|
];
|
|
1254
1290
|
var DEFAULT_VIDEO_GENERATE_MODEL = "google/gemini-omni-1.1-flash";
|
|
@@ -1828,6 +1864,67 @@ var MODEL_REGISTRY = {
|
|
|
1828
1864
|
generate_audio: { kind: "boolean" }
|
|
1829
1865
|
}
|
|
1830
1866
|
},
|
|
1867
|
+
"x-ai/grok-imagine-video-1.5": {
|
|
1868
|
+
// NOT a replacement for the entry above, and deliberately absent from
|
|
1869
|
+
// `AVATAR_VIDEO_MODELS` and `REFERENCE_TO_VIDEO_MODELS`: OpenRouter describes the
|
|
1870
|
+
// older Grok as "text-, image-, and reference-conditioned" and 1.5 as taking "an
|
|
1871
|
+
// optional starting image". Reference conditioning is the whole reason Grok carries
|
|
1872
|
+
// avatars — it is what holds a face from the identity sheet — so swapping them
|
|
1873
|
+
// would quietly cost every avatar clip its likeness.
|
|
1874
|
+
//
|
|
1875
|
+
// What it adds over the older one: 1080p, and 15s rather than 10.
|
|
1876
|
+
label: "Grok Imagine Video 1.5",
|
|
1877
|
+
inputs: [],
|
|
1878
|
+
optional_inputs: [{ kind: "image", mimes: OPENROUTER_IMAGE_MIMES }],
|
|
1879
|
+
required: ["prompt"],
|
|
1880
|
+
params: {
|
|
1881
|
+
prompt: { kind: "string" },
|
|
1882
|
+
aspect_ratio: { kind: "string", enum: ["16:9", "9:16", "1:1", "4:3", "3:4", "3:2", "2:3"] },
|
|
1883
|
+
resolution: { kind: "string", enum: ["480p", "720p", "1080p"] },
|
|
1884
|
+
duration: { kind: "number", enum: GROK_IMAGINE_15_DURATIONS }
|
|
1885
|
+
}
|
|
1886
|
+
},
|
|
1887
|
+
"alibaba/wan-3.0": {
|
|
1888
|
+
// 2.7's successor on every axis that matters: 2–30s against 2–10, 1080p and 480p
|
|
1889
|
+
// as well as 720p, and it keeps the native audio that made 2.7 the landscape
|
|
1890
|
+
// talking-head option.
|
|
1891
|
+
//
|
|
1892
|
+
// The ratios are what it ADVERTISES. 2.7 advertised five and delivered 16:9 for
|
|
1893
|
+
// all of them — measured, twice, which is why its enum here is a single value. That
|
|
1894
|
+
// was a different model and this one is not gated on its predecessor's lie, but
|
|
1895
|
+
// the first vertical clip anyone renders here is worth looking at before a campaign
|
|
1896
|
+
// is built on it.
|
|
1897
|
+
//
|
|
1898
|
+
// No `last_frame`: 2.7 takes one and 3.0 does not, so a canvas that ends a shot on
|
|
1899
|
+
// a chosen frame stays on 2.7.
|
|
1900
|
+
label: "Alibaba Wan 3.0",
|
|
1901
|
+
inputs: [],
|
|
1902
|
+
optional_inputs: [{ kind: "image", mimes: OPENROUTER_IMAGE_MIMES }],
|
|
1903
|
+
required: ["prompt"],
|
|
1904
|
+
params: {
|
|
1905
|
+
prompt: { kind: "string" },
|
|
1906
|
+
aspect_ratio: { kind: "string", enum: ["16:9", "9:16", "1:1", "4:3", "3:4"] },
|
|
1907
|
+
resolution: { kind: "string", enum: ["480p", "720p", "1080p"] },
|
|
1908
|
+
duration: { kind: "number", enum: WAN_30_DURATIONS },
|
|
1909
|
+
seed: { kind: "number" },
|
|
1910
|
+
generate_audio: { kind: "boolean" }
|
|
1911
|
+
}
|
|
1912
|
+
},
|
|
1913
|
+
"runway/gen-4.5": {
|
|
1914
|
+
// One resolution, two shapes, and silent — `generate_audio: false` in its listing,
|
|
1915
|
+
// so a talking head belongs on Wan or Omni. First frame only, no reference.
|
|
1916
|
+
label: "Runway Gen-4.5",
|
|
1917
|
+
inputs: [],
|
|
1918
|
+
optional_inputs: [{ kind: "image", mimes: OPENROUTER_IMAGE_MIMES }],
|
|
1919
|
+
required: ["prompt"],
|
|
1920
|
+
params: {
|
|
1921
|
+
prompt: { kind: "string" },
|
|
1922
|
+
aspect_ratio: { kind: "string", enum: ["16:9", "9:16"] },
|
|
1923
|
+
resolution: { kind: "string", enum: ["720p"] },
|
|
1924
|
+
duration: { kind: "number", enum: RUNWAY_45_DURATIONS },
|
|
1925
|
+
seed: { kind: "number" }
|
|
1926
|
+
}
|
|
1927
|
+
},
|
|
1831
1928
|
"alibaba/happyhorse-1.1": {
|
|
1832
1929
|
// Third reference-capable option: strong identity, correct ratios, no audio.
|
|
1833
1930
|
// First-frame only (no `last_frame`), so it cannot close a beat on a
|
|
@@ -9404,4 +9501,4 @@ export {
|
|
|
9404
9501
|
defaultRegistry,
|
|
9405
9502
|
createEngineFromEnv
|
|
9406
9503
|
};
|
|
9407
|
-
//# sourceMappingURL=chunk-
|
|
9504
|
+
//# sourceMappingURL=chunk-FENRH7OF.js.map
|