@nodaro/shared 3.63.0 → 3.65.0
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.
- package/dist/{chunk-3K24DWOH.js → chunk-222QM7CO.js} +3 -3
- package/dist/{chunk-3K24DWOH.js.map → chunk-222QM7CO.js.map} +1 -1
- package/dist/chunk-27HA6YOL.cjs +377 -0
- package/dist/chunk-27HA6YOL.cjs.map +1 -0
- package/dist/{chunk-4DVWXEB6.js → chunk-2NKF476Z.js} +3 -3
- package/dist/{chunk-4DVWXEB6.js.map → chunk-2NKF476Z.js.map} +1 -1
- package/dist/chunk-3AU52TSW.js +143 -0
- package/dist/chunk-3AU52TSW.js.map +1 -0
- package/dist/{chunk-F6WGGYQW.cjs → chunk-5NVGKXJV.cjs} +68 -29
- package/dist/chunk-5NVGKXJV.cjs.map +1 -0
- package/dist/{chunk-HWQ2SYNA.js → chunk-5SSMPYE5.js} +99 -9
- package/dist/chunk-5SSMPYE5.js.map +1 -0
- package/dist/{chunk-WRSWH7US.cjs → chunk-6BRWPJPP.cjs} +17 -2
- package/dist/chunk-6BRWPJPP.cjs.map +1 -0
- package/dist/{chunk-KHH54GRP.cjs → chunk-6ET3KDO4.cjs} +15 -338
- package/dist/chunk-6ET3KDO4.cjs.map +1 -0
- package/dist/{chunk-TM5JYSJF.cjs → chunk-6OSIOVF7.cjs} +98 -37
- package/dist/chunk-6OSIOVF7.cjs.map +1 -0
- package/dist/{chunk-YUER7TJG.js → chunk-6X5PBAC4.js} +13 -4
- package/dist/chunk-6X5PBAC4.js.map +1 -0
- package/dist/chunk-7BN4EDIC.cjs +185 -0
- package/dist/chunk-7BN4EDIC.cjs.map +1 -0
- package/dist/{chunk-6M563QE6.js → chunk-7OWXHW7Z.js} +78 -28
- package/dist/chunk-7OWXHW7Z.js.map +1 -0
- package/dist/{chunk-FXGNRYQQ.js → chunk-AT53ZQUC.js} +3 -3
- package/dist/{chunk-FXGNRYQQ.js.map → chunk-AT53ZQUC.js.map} +1 -1
- package/dist/{chunk-HSHCCBIV.js → chunk-BQYCZLMF.js} +16 -3
- package/dist/chunk-BQYCZLMF.js.map +1 -0
- package/dist/{chunk-34QLPTSA.js → chunk-BSTAAGLY.js} +3 -3
- package/dist/{chunk-34QLPTSA.js.map → chunk-BSTAAGLY.js.map} +1 -1
- package/dist/{chunk-PSKZKLDG.js → chunk-CW6MMQNO.js} +562 -26
- package/dist/chunk-CW6MMQNO.js.map +1 -0
- package/dist/{chunk-EPXGG27L.js → chunk-D5Y3EYRM.js} +5 -5
- package/dist/{chunk-EPXGG27L.js.map → chunk-D5Y3EYRM.js.map} +1 -1
- package/dist/{chunk-XGVJGNZW.cjs → chunk-DQ2PARNH.cjs} +15 -5
- package/dist/chunk-DQ2PARNH.cjs.map +1 -0
- package/dist/{chunk-6BZC2ZHC.cjs → chunk-E66HSC2E.cjs} +9 -9
- package/dist/{chunk-6BZC2ZHC.cjs.map → chunk-E66HSC2E.cjs.map} +1 -1
- package/dist/{chunk-NIAFR5EE.js → chunk-FGBGKBML.js} +32 -3
- package/dist/chunk-FGBGKBML.js.map +1 -0
- package/dist/{chunk-M5OCFGHF.cjs → chunk-FI3HJZWM.cjs} +588 -25
- package/dist/chunk-FI3HJZWM.cjs.map +1 -0
- package/dist/{chunk-L2TVY6XN.cjs → chunk-FJZTQTF4.cjs} +4 -4
- package/dist/{chunk-L2TVY6XN.cjs.map → chunk-FJZTQTF4.cjs.map} +1 -1
- package/dist/{chunk-XGGZWA4A.cjs → chunk-GJ3D5X4X.cjs} +52 -49
- package/dist/chunk-GJ3D5X4X.cjs.map +1 -0
- package/dist/{chunk-Q6IES3U2.js → chunk-H5KHWTAE.js} +5 -2
- package/dist/chunk-H5KHWTAE.js.map +1 -0
- package/dist/{chunk-MD4XHED4.js → chunk-H745DQ6V.js} +3 -3
- package/dist/{chunk-MD4XHED4.js.map → chunk-H745DQ6V.js.map} +1 -1
- package/dist/{chunk-KCZHQH2O.js → chunk-HAH2F7LT.js} +4 -4
- package/dist/{chunk-KCZHQH2O.js.map → chunk-HAH2F7LT.js.map} +1 -1
- package/dist/{chunk-V5EEQR2U.cjs → chunk-JQMFJAYU.cjs} +21 -21
- package/dist/{chunk-V5EEQR2U.cjs.map → chunk-JQMFJAYU.cjs.map} +1 -1
- package/dist/{chunk-WZVIBAKN.cjs → chunk-JYU5EIRM.cjs} +12 -12
- package/dist/{chunk-WZVIBAKN.cjs.map → chunk-JYU5EIRM.cjs.map} +1 -1
- package/dist/{chunk-Y35JQPT6.cjs → chunk-K6RIN3LW.cjs} +142 -45
- package/dist/chunk-K6RIN3LW.cjs.map +1 -0
- package/dist/{chunk-BLJUCPMV.cjs → chunk-KSHAAPQF.cjs} +5 -2
- package/dist/chunk-KSHAAPQF.cjs.map +1 -0
- package/dist/{chunk-2PU3AGGS.cjs → chunk-KTCEKGBJ.cjs} +7 -7
- package/dist/{chunk-2PU3AGGS.cjs.map → chunk-KTCEKGBJ.cjs.map} +1 -1
- package/dist/{chunk-LH72PQO2.cjs → chunk-KWYS4IIM.cjs} +165 -50
- package/dist/chunk-KWYS4IIM.cjs.map +1 -0
- package/dist/{chunk-GSKLPI3J.cjs → chunk-MIT2USTS.cjs} +118 -61
- package/dist/chunk-MIT2USTS.cjs.map +1 -0
- package/dist/{chunk-Y4GCJJPF.cjs → chunk-NYZQOIXJ.cjs} +35 -2
- package/dist/chunk-NYZQOIXJ.cjs.map +1 -0
- package/dist/{chunk-A2PX5IRC.cjs → chunk-PERUFUVH.cjs} +13 -13
- package/dist/{chunk-A2PX5IRC.cjs.map → chunk-PERUFUVH.cjs.map} +1 -1
- package/dist/{chunk-X5BAFW3Q.cjs → chunk-PGDR7CO3.cjs} +8 -8
- package/dist/{chunk-X5BAFW3Q.cjs.map → chunk-PGDR7CO3.cjs.map} +1 -1
- package/dist/{chunk-OV3JTKBY.js → chunk-PL7AQQV7.js} +123 -22
- package/dist/chunk-PL7AQQV7.js.map +1 -0
- package/dist/chunk-Q4YOLT4M.js +164 -0
- package/dist/chunk-Q4YOLT4M.js.map +1 -0
- package/dist/{chunk-GMBDLNLD.cjs → chunk-R5ILCNCC.cjs} +197 -60
- package/dist/chunk-R5ILCNCC.cjs.map +1 -0
- package/dist/chunk-RP4KQUNU.cjs +168 -0
- package/dist/chunk-RP4KQUNU.cjs.map +1 -0
- package/dist/{chunk-77WU3V3J.cjs → chunk-SPU7MUWE.cjs} +6 -6
- package/dist/{chunk-77WU3V3J.cjs.map → chunk-SPU7MUWE.cjs.map} +1 -1
- package/dist/{chunk-XBTYGQ6Z.js → chunk-ST7OBB5J.js} +8 -5
- package/dist/chunk-ST7OBB5J.js.map +1 -0
- package/dist/{chunk-2QKF7M2Z.js → chunk-TAPBDT47.js} +3 -3
- package/dist/{chunk-2QKF7M2Z.js.map → chunk-TAPBDT47.js.map} +1 -1
- package/dist/{chunk-BMJVKYVB.js → chunk-V5NLQ4L5.js} +123 -7
- package/dist/chunk-V5NLQ4L5.js.map +1 -0
- package/dist/{chunk-KMGCHV6O.js → chunk-V6OXSXIX.js} +63 -30
- package/dist/chunk-V6OXSXIX.js.map +1 -0
- package/dist/{chunk-LMAQUR55.cjs → chunk-WQMPIURW.cjs} +6 -6
- package/dist/{chunk-LMAQUR55.cjs.map → chunk-WQMPIURW.cjs.map} +1 -1
- package/dist/{chunk-ZPE2ZTJ4.cjs → chunk-WZG6GZIL.cjs} +4 -4
- package/dist/{chunk-ZPE2ZTJ4.cjs.map → chunk-WZG6GZIL.cjs.map} +1 -1
- package/dist/{chunk-AE3AAJ4O.js → chunk-XJANKTIL.js} +3 -3
- package/dist/{chunk-AE3AAJ4O.js.map → chunk-XJANKTIL.js.map} +1 -1
- package/dist/{chunk-O7RZRUMZ.js → chunk-XJWED257.js} +14 -309
- package/dist/chunk-XJWED257.js.map +1 -0
- package/dist/chunk-XSSLNINL.js +346 -0
- package/dist/chunk-XSSLNINL.js.map +1 -0
- package/dist/{chunk-6TGB7ECA.cjs → chunk-YD4XY74W.cjs} +12 -8
- package/dist/chunk-YD4XY74W.cjs.map +1 -0
- package/dist/{chunk-YX6TFX7H.js → chunk-YMZUE2IX.js} +66 -11
- package/dist/chunk-YMZUE2IX.js.map +1 -0
- package/dist/{chunk-4MMZD63H.js → chunk-Z77M7Q2R.js} +3 -3
- package/dist/{chunk-4MMZD63H.js.map → chunk-Z77M7Q2R.js.map} +1 -1
- package/dist/{chunk-CZ5IQB3L.js → chunk-ZSRXY5YT.js} +8 -5
- package/dist/chunk-ZSRXY5YT.js.map +1 -0
- package/dist/entries/apps.cjs +43 -42
- package/dist/entries/apps.js +8 -7
- package/dist/entries/captions.cjs +38 -30
- package/dist/entries/captions.d.cts +48 -2
- package/dist/entries/captions.d.ts +48 -2
- package/dist/entries/captions.js +1 -1
- package/dist/entries/credits.cjs +319 -191
- package/dist/entries/credits.d.cts +260 -19
- package/dist/entries/credits.d.ts +260 -19
- package/dist/entries/credits.js +11 -11
- package/dist/entries/entities.cjs +202 -193
- package/dist/entries/entities.d.cts +44 -9
- package/dist/entries/entities.d.ts +44 -9
- package/dist/entries/entities.js +2 -1
- package/dist/entries/llm-models.cjs +86 -62
- package/dist/entries/llm-models.d.cts +73 -20
- package/dist/entries/llm-models.d.ts +73 -20
- package/dist/entries/llm-models.js +2 -2
- package/dist/entries/media.cjs +79 -63
- package/dist/entries/media.d.cts +154 -1
- package/dist/entries/media.d.ts +154 -1
- package/dist/entries/media.js +1 -1
- package/dist/entries/model-catalog.cjs +205 -109
- package/dist/entries/model-catalog.d.cts +65 -58
- package/dist/entries/model-catalog.d.ts +65 -58
- package/dist/entries/model-catalog.js +4 -4
- package/dist/entries/models.cjs +527 -346
- package/dist/entries/models.d.cts +86 -7
- package/dist/entries/models.d.ts +86 -7
- package/dist/entries/models.js +6 -5
- package/dist/entries/nodes.cjs +220 -211
- package/dist/entries/nodes.d.cts +46 -3
- package/dist/entries/nodes.d.ts +46 -3
- package/dist/entries/nodes.js +10 -9
- package/dist/entries/pipeline.cjs +156 -152
- package/dist/entries/pipeline.d.cts +14 -4
- package/dist/entries/pipeline.d.ts +14 -4
- package/dist/entries/pipeline.js +3 -3
- package/dist/entries/social.cjs +177 -177
- package/dist/entries/social.js +2 -2
- package/dist/entries/ugc.cjs +71 -71
- package/dist/entries/ugc.js +4 -4
- package/dist/entries/video.cjs +174 -145
- package/dist/entries/video.d.cts +114 -2
- package/dist/entries/video.d.ts +114 -2
- package/dist/entries/video.js +7 -6
- package/dist/entries/workflow.cjs +150 -150
- package/dist/entries/workflow.d.cts +1 -1
- package/dist/entries/workflow.d.ts +1 -1
- package/dist/entries/workflow.js +2 -2
- package/dist/index.cjs +2430 -1929
- package/dist/index.d.cts +13 -12
- package/dist/index.d.ts +13 -12
- package/dist/index.js +28 -27
- package/dist/{model-catalog-C_Eo35tL.d.cts → model-catalog-BoUKGFNl.d.cts} +182 -3
- package/dist/{model-catalog-C_Eo35tL.d.ts → model-catalog-BoUKGFNl.d.ts} +182 -3
- package/dist/{model-constants-DWcxOmaY.d.cts → model-constants-BElS3AeE.d.ts} +142 -23
- package/dist/{model-constants-Bg4ccSQb.d.ts → model-constants-CpQz1bGa.d.cts} +142 -23
- package/dist/tts-capabilities-BtnDWK9K.d.ts +133 -0
- package/dist/tts-capabilities-C52UccTI.d.cts +133 -0
- package/package.json +1 -1
- package/src/__tests__/add-captions-nemotron-pricing.test.ts +111 -0
- package/src/__tests__/dialogue-capabilities.test.ts +167 -6
- package/src/__tests__/extend-providers.test.ts +353 -26
- package/src/__tests__/fixtures/mask-composite-quote-parity.json +1 -1
- package/src/__tests__/fixtures/measured-input-quote-parity.json +1435 -0
- package/src/__tests__/fixtures/planned-length-quote-parity.json +1866 -0
- package/src/__tests__/fixtures/video-utility-quote-parity.json +2343 -0
- package/src/__tests__/flux-3-klein.test.ts +203 -0
- package/src/__tests__/gpt-image-2-6-formats.test.ts +3 -1
- package/src/__tests__/kling4-catalog.test.ts +1 -1
- package/src/__tests__/llm-models.test.ts +48 -13
- package/src/__tests__/llm-price-key.test.ts +12 -0
- package/src/__tests__/model-makers.test.ts +3 -0
- package/src/__tests__/node-runtime-keys.test.ts +7 -0
- package/src/__tests__/normalize-node-params.test.ts +147 -1
- package/src/__tests__/retired-lip-sync-providers.test.ts +43 -0
- package/src/__tests__/seedance-extend-source-credit-id.test.ts +89 -0
- package/src/__tests__/sequence-execution.test.ts +22 -1
- package/src/__tests__/speech-pricing.test.ts +62 -1
- package/src/__tests__/suno-credit-type.test.ts +21 -2
- package/src/__tests__/transcribe-capabilities.test.ts +4 -2
- package/src/__tests__/tts-capabilities.test.ts +160 -4
- package/src/__tests__/video-analysis-catalog-sync.test.ts +4 -2
- package/src/__tests__/video-analysis-pricing.test.ts +22 -9
- package/src/__tests__/video-subjects.test.ts +192 -0
- package/src/__tests__/wired-character-names.test.ts +142 -0
- package/src/caption-styles.ts +62 -0
- package/src/credit-estimators/add-captions-nemotron.ts +73 -0
- package/src/credit-estimators/index.ts +14 -0
- package/src/credit-estimators/speech.ts +57 -9
- package/src/credit-identifiers.ts +284 -13
- package/src/dialogue-capabilities.ts +220 -3
- package/src/dubbing-structure.ts +186 -0
- package/src/entries/captions.ts +2 -0
- package/src/entries/credits.ts +25 -1
- package/src/entries/entities.ts +2 -0
- package/src/entries/llm-models.ts +7 -0
- package/src/entries/media.ts +3 -1
- package/src/entries/model-catalog.ts +30 -0
- package/src/entries/models.ts +54 -0
- package/src/entries/nodes.ts +2 -0
- package/src/entries/pipeline.ts +1 -1
- package/src/entries/video.ts +8 -1
- package/src/flux-3-klein.ts +69 -0
- package/src/gemini-tts.ts +144 -0
- package/src/input-image-pricing.ts +23 -6
- package/src/llm-models.ts +130 -36
- package/src/model-catalog.ts +539 -25
- package/src/model-constants.ts +247 -25
- package/src/node-default-mappings.ts +3 -0
- package/src/node-runtime-keys.ts +3 -0
- package/src/normalize-node-params.ts +123 -2
- package/src/seedance-extend-rate.ts +116 -0
- package/src/sequence-execution.ts +24 -0
- package/src/transcribe-preflight.ts +23 -0
- package/src/tts-capabilities.ts +193 -1
- package/src/video-analysis-pricing.ts +39 -15
- package/src/video-character-refs.ts +160 -3
- package/src/wired-character-names.ts +134 -18
- package/src/wired-character-video-subjects.ts +63 -0
- package/dist/chunk-63H4C56Q.cjs +0 -38
- package/dist/chunk-63H4C56Q.cjs.map +0 -1
- package/dist/chunk-6M563QE6.js.map +0 -1
- package/dist/chunk-6TGB7ECA.cjs.map +0 -1
- package/dist/chunk-6XUG4L7X.cjs +0 -54
- package/dist/chunk-6XUG4L7X.cjs.map +0 -1
- package/dist/chunk-BLJUCPMV.cjs.map +0 -1
- package/dist/chunk-BMJVKYVB.js.map +0 -1
- package/dist/chunk-CZ5IQB3L.js.map +0 -1
- package/dist/chunk-F6WGGYQW.cjs.map +0 -1
- package/dist/chunk-GMBDLNLD.cjs.map +0 -1
- package/dist/chunk-GSKLPI3J.cjs.map +0 -1
- package/dist/chunk-HBXF6RUU.js +0 -31
- package/dist/chunk-HBXF6RUU.js.map +0 -1
- package/dist/chunk-HSHCCBIV.js.map +0 -1
- package/dist/chunk-HWQ2SYNA.js.map +0 -1
- package/dist/chunk-IY3Y462W.js +0 -42
- package/dist/chunk-IY3Y462W.js.map +0 -1
- package/dist/chunk-KHH54GRP.cjs.map +0 -1
- package/dist/chunk-KMGCHV6O.js.map +0 -1
- package/dist/chunk-LH72PQO2.cjs.map +0 -1
- package/dist/chunk-M5OCFGHF.cjs.map +0 -1
- package/dist/chunk-NIAFR5EE.js.map +0 -1
- package/dist/chunk-O7RZRUMZ.js.map +0 -1
- package/dist/chunk-OV3JTKBY.js.map +0 -1
- package/dist/chunk-PSKZKLDG.js.map +0 -1
- package/dist/chunk-Q6IES3U2.js.map +0 -1
- package/dist/chunk-TM5JYSJF.cjs.map +0 -1
- package/dist/chunk-WRSWH7US.cjs.map +0 -1
- package/dist/chunk-XBTYGQ6Z.js.map +0 -1
- package/dist/chunk-XGGZWA4A.cjs.map +0 -1
- package/dist/chunk-XGVJGNZW.cjs.map +0 -1
- package/dist/chunk-Y35JQPT6.cjs.map +0 -1
- package/dist/chunk-Y4GCJJPF.cjs.map +0 -1
- package/dist/chunk-YUER7TJG.js.map +0 -1
- package/dist/chunk-YX6TFX7H.js.map +0 -1
package/src/model-catalog.ts
CHANGED
|
@@ -34,6 +34,20 @@ import { MINIMAX_H3_IMAGE_DEFAULT_RESOLUTION, MINIMAX_H3_IMAGE_RESOLUTIONS } fro
|
|
|
34
34
|
import { isAutoVideoDuration } from "./video-duration-auto.js"
|
|
35
35
|
import type { DeclaredSeedanceTaskType } from "./seedance-task-type.js"
|
|
36
36
|
import { SEEDREAM_5_5_DEFAULT_RESOLUTION, SEEDREAM_5_5_RESOLUTIONS } from "./seedream-5-5.js"
|
|
37
|
+
import { FLUX_3_KLEIN_RESOLUTIONS } from "./flux-3-klein.js"
|
|
38
|
+
import {
|
|
39
|
+
GEMINI_3_1_DEFAULT_VOICE,
|
|
40
|
+
GEMINI_3_1_TTS_ACCENTS,
|
|
41
|
+
GEMINI_3_1_TTS_PACES,
|
|
42
|
+
GEMINI_3_1_TTS_STYLES,
|
|
43
|
+
GEMINI_3_1_TTS_VOICES,
|
|
44
|
+
GEMINI_3_8_DEFAULT_VOICE,
|
|
45
|
+
GEMINI_3_8_TONE_TAGS,
|
|
46
|
+
GEMINI_3_8_TTS_VOICES,
|
|
47
|
+
GEMINI_TTS_MAX_CHARS,
|
|
48
|
+
TTS_DIRECTION_MAX_CHARS,
|
|
49
|
+
type TtsNamedVoice,
|
|
50
|
+
} from "./gemini-tts.js"
|
|
37
51
|
|
|
38
52
|
export type ModelKind = "image" | "video" | "audio"
|
|
39
53
|
|
|
@@ -77,7 +91,14 @@ export type ModelMode =
|
|
|
77
91
|
| "video-audit"
|
|
78
92
|
|
|
79
93
|
export interface PriceVariant {
|
|
80
|
-
/**
|
|
94
|
+
/**
|
|
95
|
+
* Composite identifier as it appears in `STATIC_CREDIT_COSTS` — or, for a
|
|
96
|
+
* model a run is charged a multiple of a rate for, the whole-run quote id of
|
|
97
|
+
* one run (`<rateId>:<units>u`, `unitQuoteCreditId`): its rate's row × the
|
|
98
|
+
* units, rounded up once, as the price table and `GET
|
|
99
|
+
* /v1/credits/model-cost` price it (Seedance 2.5 Extend lists its default
|
|
100
|
+
* run this way). `credits` is then that run's list price.
|
|
101
|
+
*/
|
|
81
102
|
identifier: string
|
|
82
103
|
/** The catalog's list price in credits. `GET /v1/models` and the MCP
|
|
83
104
|
* `list_models` tool serve the price a run is charged on the instance,
|
|
@@ -100,6 +121,14 @@ export interface PriceVariant {
|
|
|
100
121
|
*/
|
|
101
122
|
export const SPEECH_UNIT_PRICE_NOTE = "per started 100 characters, at least 8 units — where length pricing is on; a request costs at most its units × this"
|
|
102
123
|
|
|
124
|
+
/**
|
|
125
|
+
* The note on a speech model's `<model>:per-1k-chars` row (`TtsCapabilities.ratePer1kChars`):
|
|
126
|
+
* the same count as {@link SPEECH_UNIT_PRICE_NOTE} — started 100 characters, at
|
|
127
|
+
* least 8 units — quoted per 1,000, so a request costs its hundreds × this ÷ 10,
|
|
128
|
+
* rounded up once (the exact price of N hundreds is the `<model>:per-100-chars:<N>u` id).
|
|
129
|
+
*/
|
|
130
|
+
export const SPEECH_PER_1K_PRICE_NOTE = "per 1,000 characters, counted in started 100s, at least 800 — where length pricing is on; a request costs its hundreds × this ÷ 10, rounded up"
|
|
131
|
+
|
|
103
132
|
/** A voice-setting lever a speech model honours. */
|
|
104
133
|
export type TtsSettingLever = "stability" | "similarity" | "style" | "speed" | "speakerBoost"
|
|
105
134
|
|
|
@@ -152,6 +181,81 @@ export interface TtsCapabilities {
|
|
|
152
181
|
maxChars: number
|
|
153
182
|
/** Languages offered in pickers (ISO 639-1; `fil` has no two-letter code). A curated list, not the vendor's full one. */
|
|
154
183
|
languages: readonly string[]
|
|
184
|
+
/**
|
|
185
|
+
* The model speaks only these named voices, by wire name: a request's
|
|
186
|
+
* `voice` must be one of them (else a 400 before any hold), and none means
|
|
187
|
+
* `defaultVoice`. Absent ⇒ the ElevenLabs voices (premade, library, clones).
|
|
188
|
+
* A client offers this list in place of the ElevenLabs voice browser.
|
|
189
|
+
*/
|
|
190
|
+
voices?: readonly TtsNamedVoice[]
|
|
191
|
+
/** With `voices`: the voice a request that names none runs with. */
|
|
192
|
+
defaultVoice?: string
|
|
193
|
+
/**
|
|
194
|
+
* With `audioTags`: the tags are written `<in angle brackets>`, from the
|
|
195
|
+
* closed list `tags`. The platform's `[square]` tags are rewritten to them
|
|
196
|
+
* where they match and the rest are taken out before sending
|
|
197
|
+
* (`@nodaro/prompts` `toAngleTags`). Absent ⇒ `[square]` tags, sent as written.
|
|
198
|
+
*/
|
|
199
|
+
tagSyntax?: "angle"
|
|
200
|
+
/** With `tagSyntax: "angle"`: the tags the model performs, by their own names (`laugh`, `short pause`). */
|
|
201
|
+
tags?: readonly string[]
|
|
202
|
+
/**
|
|
203
|
+
* Style direction the model takes beside the text — never spoken, never
|
|
204
|
+
* billed. Absent ⇒ none (the request's direction fields are not sent).
|
|
205
|
+
*/
|
|
206
|
+
direction?: TtsDirectionSheet
|
|
207
|
+
/**
|
|
208
|
+
* Dialogue lane only: at most this many speakers (distinct voices) in one
|
|
209
|
+
* script. Absent ⇒ the lane's own limit (ElevenLabs dialogue: 10).
|
|
210
|
+
*/
|
|
211
|
+
dialogueSpeakers?: number
|
|
212
|
+
/**
|
|
213
|
+
* Dialogue lane only: the model can add natural filler words ("um", "you
|
|
214
|
+
* know") between two speakers when asked (`fillerWords` on a request).
|
|
215
|
+
*/
|
|
216
|
+
fillerWords?: true
|
|
217
|
+
/**
|
|
218
|
+
* Dialogue lane only: the model delivers every speech file format
|
|
219
|
+
* (`TTS_OUTPUT_FORMATS`) — it renders a WAV the platform keeps as it is or
|
|
220
|
+
* encodes to MP3. Absent ⇒ the default MP3 only (the ElevenLabs dialogue
|
|
221
|
+
* endpoint, as the platform calls it).
|
|
222
|
+
*/
|
|
223
|
+
dialogueAnyOutputFormat?: true
|
|
224
|
+
/**
|
|
225
|
+
* The model clones a speaker's voice for one request from two recordings
|
|
226
|
+
* (`replica`: a reference and a consent recording, `TtsReplica`), in place of
|
|
227
|
+
* a named voice; nothing is saved as a voice. Gemini 3.8 Flash only: KIE's
|
|
228
|
+
* 3.8 Flash-Lite refuses it ("Inline voice sample replication is not
|
|
229
|
+
* supported", probed 2026-11-13).
|
|
230
|
+
*/
|
|
231
|
+
replica?: true
|
|
232
|
+
/**
|
|
233
|
+
* The model's length-price row is quoted per 1,000 characters
|
|
234
|
+
* (`<id>:per-1k-chars`) instead of per started 100 (`<id>:per-100-chars`).
|
|
235
|
+
* The count is the same — started 100 characters, at least 8 — and a
|
|
236
|
+
* request costs its hundreds × the row ÷ 10, rounded up once
|
|
237
|
+
* (`speechRowCredits`). For a rate below one credit per 100 characters,
|
|
238
|
+
* which a whole-credit row cannot hold.
|
|
239
|
+
*/
|
|
240
|
+
ratePer1kChars?: true
|
|
241
|
+
}
|
|
242
|
+
|
|
243
|
+
/**
|
|
244
|
+
* The style direction a speech model takes (`styleDirection`, `pace`,
|
|
245
|
+
* `accent`, `audioProfile` on a request). A field the sheet does not name is
|
|
246
|
+
* never sent.
|
|
247
|
+
*/
|
|
248
|
+
export interface TtsDirectionSheet {
|
|
249
|
+
/** `styleDirection` is free text — tone, emotion, pace in the person's own words — up to this many characters. */
|
|
250
|
+
styleMaxChars?: number
|
|
251
|
+
/** `styleDirection` is one of these (the model's own names). */
|
|
252
|
+
styles?: readonly string[]
|
|
253
|
+
/** `pace` is one of these. */
|
|
254
|
+
paces?: readonly string[]
|
|
255
|
+
/** `accent` is one of these. */
|
|
256
|
+
accents?: readonly string[]
|
|
257
|
+
/** `audioProfile` (who is speaking, in words) is free text up to this many characters. */
|
|
258
|
+
audioProfileMaxChars?: number
|
|
155
259
|
}
|
|
156
260
|
|
|
157
261
|
/**
|
|
@@ -173,6 +277,7 @@ export const MODEL_MAKERS = {
|
|
|
173
277
|
bytedance: { name: "ByteDance" },
|
|
174
278
|
elevenlabs: { name: "ElevenLabs" },
|
|
175
279
|
google: { name: "Google" },
|
|
280
|
+
heygen: { name: "HeyGen" },
|
|
176
281
|
ideogram: { name: "Ideogram" },
|
|
177
282
|
kuaishou: { name: "Kuaishou" },
|
|
178
283
|
lightricks: { name: "Lightricks" },
|
|
@@ -339,7 +444,10 @@ export interface ModelCatalogEntryFields {
|
|
|
339
444
|
* marked required, with a closed list (Seedream 5.5: text to image requires
|
|
340
445
|
* one of its eight ratios and has no "auto"; its edit takes "auto" and the
|
|
341
446
|
* same eight; docs.kie.ai/market/seedream/5-5-text-to-image.md and
|
|
342
|
-
* 5-5-image-to-image.md).
|
|
447
|
+
* 5-5-image-to-image.md). FLUX 3 Klein's schemas give the same evidence: a
|
|
448
|
+
* closed list of nine ratios on text to image, with no "auto", and the same
|
|
449
|
+
* nine plus "auto" on the edit (docs.kie.ai/market/flux3/klein-text-to-image.md
|
|
450
|
+
* and klein-edit-image.md). The lanes that pass a stored ratio straight on (a
|
|
343
451
|
* Scene, a surround continuation) move a ratio such a model does not list to
|
|
344
452
|
* the nearest one it does, checked against the model the request reaches
|
|
345
453
|
* (`aspectRatioForStrictModel` in the backend), and send every other model's
|
|
@@ -593,6 +701,7 @@ export const MODEL_RECOMMENDATIONS: readonly ModelRecommendation[] = [
|
|
|
593
701
|
{ intent: "video with start + end frame", modelIds: ["veo3", "veo3.1", "kling-turbo", "minimax", "hailuo-standard", "seedance-2"], note: "All listed support an end frame; VEO uses imageUrls[start, end]." },
|
|
594
702
|
{ intent: "music / song generation", modelIds: ["suno-v6", "suno-v6_wild", "suno-v6_mini", "suno-v5_5"], note: "V6 is the default flagship; V6 Wild for bolder, less predictable results; V6 Mini when speed matters; v5.5 / v5 / v4 keep their own character. Same price." },
|
|
595
703
|
{ intent: "voice over / narration", modelIds: ["elevenlabs-v4", "elevenlabs-v3", "elevenlabs-v4-turbo", "elevenlabs-turbo"], note: "v4 is the default: [audio tags] for emotion and up to 10,000 characters per request. v3 is the previous expressive model, still selectable. v4 Turbo keeps v4's tags at Turbo's credit price; Turbo v2.5 is the cheapest for plain narration without tags." },
|
|
704
|
+
{ intent: "low-cost narration with a named voice", modelIds: ["gemini-3-8-flash-lite-tts", "gemini-3-8-flash-tts", "gemini-3-1-flash-tts", "gemini-2-5-pro-tts"], note: "Google's Gemini voices cost far less per character than ElevenLabs. Each speaks only its own named voices (no clones or library voices) and no language is set: the text decides. 3.8 Flash-Lite is the cheapest; 3.8 Flash is the larger 3.8 model; both take their own tone tags (laugh, sigh, short pause…) and a free-text style. 3.1 Flash and 2.5 Pro take [audio tags] and a style, pace, accent and audio profile." },
|
|
596
705
|
{ intent: "lip-sync a portrait to audio", modelIds: ["kling-avatar-pro", "kling-avatar", "infinitalk"], note: "Pro for best mouth shape; InfiniTalk for resolution control." },
|
|
597
706
|
{ intent: "transcription / captions", modelIds: ["elevenlabs-stt", "incredibly-fast-whisper", "whisper", "nemotron-3-5-transcribe"], note: "Captions need WORD timestamps: ElevenLabs STT (always), Incredibly Fast Whisper or Nemotron 3.5 Transcribe (on request, with a speaker per word). Plain Whisper returns phrase segments only. To translate the speech (to or from English) instead of transcribing it, use Nemotron 3.5 Transcribe with a target language; a translation carries no word timings." },
|
|
598
707
|
{ intent: "motion transfer (drive a subject by another video)", modelIds: ["motion-transfer", "kling-3.0-motion"], note: "Kling 2.6 base is cheap; Kling 3.0 is premium." },
|
|
@@ -711,6 +820,14 @@ const SEEDREAM_5_5_I2I_RATIOS = ["auto", ...SEEDREAM_RATIOS] as const
|
|
|
711
820
|
// moves it to auto with JPEG), and PNG (its default) or JPEG.
|
|
712
821
|
const SEEDREAM_5_5_BACKGROUNDS = ["auto", "opaque", "transparent"] as const
|
|
713
822
|
const SEEDREAM_5_5_FORMATS = ["png", "jpeg"] as const
|
|
823
|
+
// FLUX 3 Klein (docs.kie.ai/market/flux3/klein-text-to-image + klein-edit-image):
|
|
824
|
+
// both endpoints list the same nine ratios, 1:1 first (the text-to-image
|
|
825
|
+
// default). The edit adds "auto", its own default (the first image's shape),
|
|
826
|
+
// listed LAST so a fallback that takes the first ratio lands on 1:1.
|
|
827
|
+
const FLUX_3_KLEIN_RATIOS = ["1:1", "4:3", "3:4", "3:2", "2:3", "16:9", "9:16", "21:9", "9:21"] as const
|
|
828
|
+
const FLUX_3_KLEIN_I2I_RATIOS = [...FLUX_3_KLEIN_RATIOS, "auto"] as const
|
|
829
|
+
// Its files: PNG (its default) or JPEG, the same price (decided 2026-11-12).
|
|
830
|
+
const FLUX_3_KLEIN_FORMATS = ["png", "jpeg"] as const
|
|
714
831
|
const Z_IMAGE_RATIOS = ["1:1", "16:9", "9:16", "4:3", "3:4"] as const
|
|
715
832
|
|
|
716
833
|
const VIDEO_RATIOS_HV = ["16:9", "9:16"] as const
|
|
@@ -963,6 +1080,63 @@ const IMAGE_MODELS: Record<string, ModelCatalogEntry> = {
|
|
|
963
1080
|
aspectRatios: KONTEXT_RATIOS,
|
|
964
1081
|
pricing: [{ identifier: "flux-kontext-max", credits: 25 }],
|
|
965
1082
|
},
|
|
1083
|
+
// ── FLUX 3 Klein ──
|
|
1084
|
+
// The light model of Black Forest Labs' FLUX 3 (KIE `flux-3/klein-text-to-image`
|
|
1085
|
+
// and `flux-3/klein-edit-image`). A different model from FLUX 2 Klein
|
|
1086
|
+
// (`flux-2-klein`, below). A prompt runs the text-to-image id; with reference
|
|
1087
|
+
// images the run moves to the edit id (`T2I_TO_I2I_VARIANT`). Priced by size
|
|
1088
|
+
// alone, the same on both ids: reference images cost nothing
|
|
1089
|
+
// (`flux3KleinCreditId`). Both schemas list a closed set of ratios and text to
|
|
1090
|
+
// image has no "auto", so both entries declare `strictAspectRatios`. PNG (the
|
|
1091
|
+
// default) or JPEG, as its siblings write (decided 2026-11-12); KIE's WebP is
|
|
1092
|
+
// not offered. KIE's prompt enhancement is left on (never sent), and its
|
|
1093
|
+
// custom width/height and several pictures per request are not offered.
|
|
1094
|
+
// See: docs.kie.ai/market/flux3/klein-text-to-image.md
|
|
1095
|
+
// docs.kie.ai/market/flux3/klein-edit-image.md
|
|
1096
|
+
"flux-3-klein": {
|
|
1097
|
+
id: "flux-3-klein",
|
|
1098
|
+
kind: "image",
|
|
1099
|
+
modes: ["t2i"] as const,
|
|
1100
|
+
family: "Black Forest Labs",
|
|
1101
|
+
maker: "black-forest-labs",
|
|
1102
|
+
label: "FLUX 3 Klein",
|
|
1103
|
+
series: "Flux",
|
|
1104
|
+
description: "FLUX 3 Klein - the light model of FLUX 3: fast and low cost, at 1K by default, 2K or 4K, in nine shapes, PNG or JPEG. Add images and it edits from up to 10 references, at the same price.",
|
|
1105
|
+
useCases: ["fast", "drafts", "iteration", "realistic"],
|
|
1106
|
+
features: ["reference-image"],
|
|
1107
|
+
aspectRatios: FLUX_3_KLEIN_RATIOS,
|
|
1108
|
+
strictAspectRatios: true,
|
|
1109
|
+
resolutions: FLUX_3_KLEIN_RESOLUTIONS,
|
|
1110
|
+
outputFormats: FLUX_3_KLEIN_FORMATS,
|
|
1111
|
+
pricing: [
|
|
1112
|
+
{ identifier: "flux-3-klein", credits: 4, note: "1K default" },
|
|
1113
|
+
{ identifier: "flux-3-klein:1K", credits: 4, note: "1K" },
|
|
1114
|
+
{ identifier: "flux-3-klein:2K", credits: 7, note: "2K" },
|
|
1115
|
+
{ identifier: "flux-3-klein:4K", credits: 17, note: "4K" },
|
|
1116
|
+
],
|
|
1117
|
+
},
|
|
1118
|
+
"flux-3-klein-i2i": {
|
|
1119
|
+
id: "flux-3-klein-i2i",
|
|
1120
|
+
kind: "image",
|
|
1121
|
+
modes: ["i2i"] as const,
|
|
1122
|
+
family: "Black Forest Labs",
|
|
1123
|
+
maker: "black-forest-labs",
|
|
1124
|
+
label: "FLUX 3 Klein (I2I)",
|
|
1125
|
+
series: "Flux",
|
|
1126
|
+
description: "FLUX 3 Klein edits - fast, low-cost changes from 1 to 10 reference images, which cost nothing extra. 1K by default, 2K or 4K, PNG or JPEG; auto keeps the first image's shape.",
|
|
1127
|
+
useCases: ["edit", "iteration", "multi-reference", "fast"],
|
|
1128
|
+
features: ["reference-image"],
|
|
1129
|
+
aspectRatios: FLUX_3_KLEIN_I2I_RATIOS,
|
|
1130
|
+
strictAspectRatios: true,
|
|
1131
|
+
resolutions: FLUX_3_KLEIN_RESOLUTIONS,
|
|
1132
|
+
outputFormats: FLUX_3_KLEIN_FORMATS,
|
|
1133
|
+
pricing: [
|
|
1134
|
+
{ identifier: "flux-3-klein-i2i", credits: 4, note: "1K default" },
|
|
1135
|
+
{ identifier: "flux-3-klein-i2i:1K", credits: 4, note: "1K" },
|
|
1136
|
+
{ identifier: "flux-3-klein-i2i:2K", credits: 7, note: "2K" },
|
|
1137
|
+
{ identifier: "flux-3-klein-i2i:4K", credits: 17, note: "4K" },
|
|
1138
|
+
],
|
|
1139
|
+
},
|
|
966
1140
|
// ── Replicate "Open" (uncensored) — run direct via Replicate, not KIE ──
|
|
967
1141
|
"flux-2-klein": {
|
|
968
1142
|
id: "flux-2-klein",
|
|
@@ -3428,8 +3602,11 @@ const VIDEO_MODELS: Record<string, ModelCatalogEntry> = {
|
|
|
3428
3602
|
useCases: ["extend"],
|
|
3429
3603
|
pricing: [{ identifier: "runway-extend", credits: 320 }],
|
|
3430
3604
|
},
|
|
3431
|
-
// Seedance extends ANY video, one provider per model.
|
|
3432
|
-
//
|
|
3605
|
+
// Seedance extends ANY video, one provider per model. Seedance 2.5 is charged
|
|
3606
|
+
// per second of video sent (the source's last 10 s, all of a shorter one)
|
|
3607
|
+
// and added, at its resolution's row, the price of two seconds (decided
|
|
3608
|
+
// 2026-11-12). Seedance 2.0 is priced on its own rows
|
|
3609
|
+
// (`tailExtendCreditId`), one per second added and resolution.
|
|
3433
3610
|
// The earlier single id `seedance-2-extend` is retired and runs as Seedance 2.5.
|
|
3434
3611
|
"seedance-2-5-extend": {
|
|
3435
3612
|
id: "seedance-2-5-extend",
|
|
@@ -3439,15 +3616,18 @@ const VIDEO_MODELS: Record<string, ModelCatalogEntry> = {
|
|
|
3439
3616
|
maker: "bytedance",
|
|
3440
3617
|
label: "Seedance 2.5 Extend",
|
|
3441
3618
|
series: "Seedance",
|
|
3442
|
-
description: "Extend ANY video with Seedance 2.5: 4–30s continuation (audio included),
|
|
3619
|
+
description: "Extend ANY video with Seedance 2.5: a 4–30s continuation (audio included) of the source's last 10 seconds, joined onto it. Charged per second of video sent and added; the figure is the default run, 8s added at 720p to a source of 10s or more.",
|
|
3443
3620
|
useCases: ["extend"],
|
|
3444
3621
|
durations: [4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30],
|
|
3445
3622
|
defaultDuration: 8,
|
|
3446
3623
|
resolutions: ["480p", "720p", "1080p"],
|
|
3447
3624
|
pricing: [
|
|
3448
|
-
|
|
3449
|
-
|
|
3450
|
-
|
|
3625
|
+
// The default run (8 s added to a source of 10 s or more: 18 s sent and
|
|
3626
|
+
// added, 9 units of the two-second row) at each resolution. A run is
|
|
3627
|
+
// charged its own seconds: `seedance-2-5-extend:<res>:per-2s` × (sent + added) ÷ 2.
|
|
3628
|
+
{ identifier: "seedance-2-5-extend:720p:per-2s:9u", credits: 1710, note: "default run: 8s added at 720p to a source of 10s or more — priced per second sent and added" },
|
|
3629
|
+
{ identifier: "seedance-2-5-extend:480p:per-2s:9u", credits: 765, note: "8s added at 480p to a source of 10s or more" },
|
|
3630
|
+
{ identifier: "seedance-2-5-extend:1080p:per-2s:9u", credits: 4275, note: "8s added at 1080p to a source of 10s or more" },
|
|
3451
3631
|
],
|
|
3452
3632
|
},
|
|
3453
3633
|
"seedance-2-0-extend": {
|
|
@@ -3492,6 +3672,63 @@ const VIDEO_MODELS: Record<string, ModelCatalogEntry> = {
|
|
|
3492
3672
|
{ identifier: "kling-4-extend:5s:1080p", credits: 638, note: "5s 1080p" },
|
|
3493
3673
|
],
|
|
3494
3674
|
},
|
|
3675
|
+
// PixVerse V6 extends ANY video natively (no tail recipe): the whole source
|
|
3676
|
+
// goes to its extend, which returns it with the new seconds joined on; the
|
|
3677
|
+
// run keeps the user's own source and joins only the new seconds to it. Its
|
|
3678
|
+
// own entry, so a deployment can withhold it apart from PixVerse V6 Max, and
|
|
3679
|
+
// its own rows (`pixverseExtendCreditId`): one per second added and quality,
|
|
3680
|
+
// `:audio` with sound (on unless turned off). `resolutions` is ascending
|
|
3681
|
+
// (repo convention); the default quality (720p) is the extend's own, as on
|
|
3682
|
+
// the other extend entries (`pixverseExtendQuality`), never declared here.
|
|
3683
|
+
"pixverse-v6-extend": {
|
|
3684
|
+
id: "pixverse-v6-extend",
|
|
3685
|
+
kind: "video",
|
|
3686
|
+
modes: ["extend"] as const,
|
|
3687
|
+
family: "PixVerse",
|
|
3688
|
+
maker: "pixverse",
|
|
3689
|
+
label: "PixVerse V6 Extend",
|
|
3690
|
+
series: "PixVerse",
|
|
3691
|
+
description: "Extend ANY video with PixVerse V6: 1–15s continuation at 360p to 1080p, with optional sound, joined onto your original clip.",
|
|
3692
|
+
useCases: ["extend"],
|
|
3693
|
+
durations: [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15],
|
|
3694
|
+
defaultDuration: 5,
|
|
3695
|
+
resolutions: ["360p", "540p", "720p", "1080p"],
|
|
3696
|
+
pricing: [
|
|
3697
|
+
{ identifier: "pixverse-v6-extend:5s:720p:audio", credits: 120, note: "default 5s 720p with sound — see :Ns:quality and :Ns:quality:audio variants for exact" },
|
|
3698
|
+
{ identifier: "pixverse-v6-extend:5s:720p", credits: 90, note: "5s 720p, no sound" },
|
|
3699
|
+
{ identifier: "pixverse-v6-extend:5s:540p:audio", credits: 90, note: "5s 540p with sound" },
|
|
3700
|
+
{ identifier: "pixverse-v6-extend:1s:360p", credits: 10, note: "1s 360p, no sound (min)" },
|
|
3701
|
+
{ identifier: "pixverse-v6-extend:15s:1080p:audio", credits: 690, note: "15s 1080p with sound (max)" },
|
|
3702
|
+
],
|
|
3703
|
+
},
|
|
3704
|
+
// Grok extends only its OWN clips: a completed Grok Imagine text-to-video
|
|
3705
|
+
// or image-to-video run of the same user, named by its task (KIE takes no
|
|
3706
|
+
// clip URL and no other model's task). It adds 6 or 10 s and returns the
|
|
3707
|
+
// source with them joined on, delivered as it is. Billed at the SOURCE's
|
|
3708
|
+
// resolution, which is never a lever — so no `resolutions` here: the run is
|
|
3709
|
+
// priced on the row the source check answers (`grokExtendCreditId`), and a
|
|
3710
|
+
// run whose source is not known yet is quoted at the dearer 720p row, the
|
|
3711
|
+
// row this entry leads with. Its own entry, so a deployment can withhold it
|
|
3712
|
+
// apart from Grok Imagine.
|
|
3713
|
+
"grok-extend": {
|
|
3714
|
+
id: "grok-extend",
|
|
3715
|
+
kind: "video",
|
|
3716
|
+
modes: ["extend"] as const,
|
|
3717
|
+
family: "xAI",
|
|
3718
|
+
maker: "xai",
|
|
3719
|
+
label: "Grok Imagine Extend",
|
|
3720
|
+
series: "Grok",
|
|
3721
|
+
description: "Extend a Grok Imagine video you made here (from text or an image) by 6 or 10 seconds, at the video's own resolution (480p or 720p).",
|
|
3722
|
+
useCases: ["extend"],
|
|
3723
|
+
durations: [6, 10],
|
|
3724
|
+
defaultDuration: 6,
|
|
3725
|
+
pricing: [
|
|
3726
|
+
{ identifier: "grok-extend:6s:720p", credits: 70, note: "6s from a 720p source (the quote while the source is not known)" },
|
|
3727
|
+
{ identifier: "grok-extend:6s:480p", credits: 40, note: "6s from a 480p source" },
|
|
3728
|
+
{ identifier: "grok-extend:10s:480p", credits: 60, note: "10s from a 480p source" },
|
|
3729
|
+
{ identifier: "grok-extend:10s:720p", credits: 120, note: "10s from a 720p source" },
|
|
3730
|
+
],
|
|
3731
|
+
},
|
|
3495
3732
|
"veo-1080p": {
|
|
3496
3733
|
id: "veo-1080p",
|
|
3497
3734
|
kind: "video",
|
|
@@ -3798,6 +4035,61 @@ const VIDEO_MODELS: Record<string, ModelCatalogEntry> = {
|
|
|
3798
4035
|
{ identifier: "volcengine-lipsync:300s", credits: 6000, note: "5-min ceiling" },
|
|
3799
4036
|
],
|
|
3800
4037
|
},
|
|
4038
|
+
// LatentSync, HeyGen Lipsync Precision and Lipsync 2 Pro run on Replicate
|
|
4039
|
+
// (providers/replicate/lip-sync.ts). Listed so a deployment's model
|
|
4040
|
+
// restriction can name them (decided 2026-11-12); what they are sent and
|
|
4041
|
+
// charged did not change. They take no size: the result keeps the size of the
|
|
4042
|
+
// video they are given.
|
|
4043
|
+
"latentsync": {
|
|
4044
|
+
id: "latentsync",
|
|
4045
|
+
kind: "video",
|
|
4046
|
+
modes: ["lip-sync"] as const,
|
|
4047
|
+
family: "LatentSync",
|
|
4048
|
+
maker: "bytedance",
|
|
4049
|
+
label: "LatentSync",
|
|
4050
|
+
series: "LatentSync",
|
|
4051
|
+
description: "Diffusion lip sync on an existing video — strong on singing and expressive vocals. Video input.",
|
|
4052
|
+
useCases: ["lip-sync", "dubbing", "video-to-video"],
|
|
4053
|
+
pricing: [{ identifier: "latentsync", credits: 7 }],
|
|
4054
|
+
},
|
|
4055
|
+
"heygen-lipsync-precision": {
|
|
4056
|
+
id: "heygen-lipsync-precision",
|
|
4057
|
+
kind: "video",
|
|
4058
|
+
modes: ["lip-sync"] as const,
|
|
4059
|
+
family: "HeyGen",
|
|
4060
|
+
maker: "heygen",
|
|
4061
|
+
label: "HeyGen Lipsync Precision",
|
|
4062
|
+
series: "HeyGen",
|
|
4063
|
+
description: "Dub existing footage with precise lip sync to a new audio track. Video input, billed per second.",
|
|
4064
|
+
useCases: ["lip-sync", "dubbing", "video-to-video"],
|
|
4065
|
+
pricing: [
|
|
4066
|
+
{ identifier: "heygen-lipsync-precision", credits: 10010, note: "5-min ceiling (no duration given)" },
|
|
4067
|
+
{ identifier: "heygen-lipsync-precision:15s", credits: 510 },
|
|
4068
|
+
{ identifier: "heygen-lipsync-precision:30s", credits: 1010 },
|
|
4069
|
+
{ identifier: "heygen-lipsync-precision:60s", credits: 2010 },
|
|
4070
|
+
{ identifier: "heygen-lipsync-precision:120s", credits: 4010 },
|
|
4071
|
+
{ identifier: "heygen-lipsync-precision:300s", credits: 10010, note: "5-min ceiling" },
|
|
4072
|
+
],
|
|
4073
|
+
},
|
|
4074
|
+
"lipsync-2-pro": {
|
|
4075
|
+
id: "lipsync-2-pro",
|
|
4076
|
+
kind: "video",
|
|
4077
|
+
modes: ["lip-sync"] as const,
|
|
4078
|
+
family: "Sync",
|
|
4079
|
+
maker: "sync-labs",
|
|
4080
|
+
label: "Lipsync 2 Pro",
|
|
4081
|
+
series: "Sync",
|
|
4082
|
+
description: "Studio-grade lip sync on existing footage, re-synced to a new audio track. Video input, billed per second.",
|
|
4083
|
+
useCases: ["lip-sync", "dubbing", "video-to-video", "premium"],
|
|
4084
|
+
pricing: [
|
|
4085
|
+
{ identifier: "lipsync-2-pro", credits: 12490, note: "5-min ceiling (no duration given)" },
|
|
4086
|
+
{ identifier: "lipsync-2-pro:15s", credits: 630 },
|
|
4087
|
+
{ identifier: "lipsync-2-pro:30s", credits: 1250 },
|
|
4088
|
+
{ identifier: "lipsync-2-pro:60s", credits: 2500 },
|
|
4089
|
+
{ identifier: "lipsync-2-pro:120s", credits: 5000 },
|
|
4090
|
+
{ identifier: "lipsync-2-pro:300s", credits: 12490, note: "5-min ceiling" },
|
|
4091
|
+
],
|
|
4092
|
+
},
|
|
3801
4093
|
// ── Video Depth ── (video-depth node; priced by the input clip's length via
|
|
3802
4094
|
// videoDepthCreditId in video-depth.ts — `<provider>:<n>s`, one row per whole
|
|
3803
4095
|
// second of the clip (`pricedClipSeconds`), representative rows below: `:1s` …
|
|
@@ -3843,8 +4135,8 @@ const VIDEO_MODELS: Record<string, ModelCatalogEntry> = {
|
|
|
3843
4135
|
// structural credit formula lives in video-analysis-pricing.ts. Composite id
|
|
3844
4136
|
// shape `video-analysis:<model>:<bucket>s` mirrors volcengine-lipsync: bare
|
|
3845
4137
|
// row = per-model namespace root at the 10-min ceiling, plus one row per bucket.)
|
|
3846
|
-
"gemini-3-flash-video-analysis": {
|
|
3847
|
-
id: "gemini-3-flash-video-analysis",
|
|
4138
|
+
"gemini-3.8-flash-video-analysis": {
|
|
4139
|
+
id: "gemini-3.8-flash-video-analysis",
|
|
3848
4140
|
kind: "video",
|
|
3849
4141
|
modes: ["video-analysis"] as const,
|
|
3850
4142
|
family: "Nodaro",
|
|
@@ -3853,6 +4145,24 @@ const VIDEO_MODELS: Record<string, ModelCatalogEntry> = {
|
|
|
3853
4145
|
series: "Video Analysis",
|
|
3854
4146
|
description: "Analyze a video into a structured shot list (scenes, camera, audio) — fast, economy tier. Billed per duration bucket.",
|
|
3855
4147
|
useCases: ["video-analysis", "shot-list", "fast"],
|
|
4148
|
+
pricing: [
|
|
4149
|
+
{ identifier: "video-analysis:gemini-3.8-flash", credits: 880, note: "10-min ceiling (no duration given)" },
|
|
4150
|
+
{ identifier: "video-analysis:gemini-3.8-flash:60s", credits: 186 },
|
|
4151
|
+
{ identifier: "video-analysis:gemini-3.8-flash:180s", credits: 193 },
|
|
4152
|
+
{ identifier: "video-analysis:gemini-3.8-flash:360s", credits: 534 },
|
|
4153
|
+
{ identifier: "video-analysis:gemini-3.8-flash:600s", credits: 880, note: "10-min ceiling" },
|
|
4154
|
+
],
|
|
4155
|
+
},
|
|
4156
|
+
"gemini-3-flash-video-analysis": {
|
|
4157
|
+
id: "gemini-3-flash-video-analysis",
|
|
4158
|
+
kind: "video",
|
|
4159
|
+
modes: ["video-analysis"] as const,
|
|
4160
|
+
family: "Nodaro",
|
|
4161
|
+
maker: "nodaro",
|
|
4162
|
+
label: "Video Analysis (Fast — previous)",
|
|
4163
|
+
series: "Video Analysis",
|
|
4164
|
+
description: "Previous fast-tier analysis model (backed the fast tier until 2026-11-13). Kept so stored raw-model configs keep running and keep pricing under their own identifier; their runs are served by the current fast model.",
|
|
4165
|
+
useCases: ["video-analysis", "shot-list", "fast"],
|
|
3856
4166
|
pricing: [
|
|
3857
4167
|
{ identifier: "video-analysis:gemini-3-flash", credits: 848, note: "10-min ceiling (no duration given)" },
|
|
3858
4168
|
{ identifier: "video-analysis:gemini-3-flash:60s", credits: 181 },
|
|
@@ -3897,6 +4207,29 @@ const VIDEO_MODELS: Record<string, ModelCatalogEntry> = {
|
|
|
3897
4207
|
{ identifier: "video-analysis:gemini-3.1-pro:600s", credits: 1056, note: "10-min ceiling" },
|
|
3898
4208
|
],
|
|
3899
4209
|
},
|
|
4210
|
+
// ULTRA (since 2026-11-13) — the deepest-reasoning analysis model reading
|
|
4211
|
+
// every window, at the same frame sampling as the other single-model tiers.
|
|
4212
|
+
// Priced between the mixed tiers and smart.
|
|
4213
|
+
"gemini-4-carbon-video-analysis": {
|
|
4214
|
+
id: "gemini-4-carbon-video-analysis",
|
|
4215
|
+
kind: "video",
|
|
4216
|
+
modes: ["video-analysis"] as const,
|
|
4217
|
+
family: "Nodaro",
|
|
4218
|
+
maker: "nodaro",
|
|
4219
|
+
label: "Video Analysis (Ultra)",
|
|
4220
|
+
series: "Video Analysis",
|
|
4221
|
+
description: "Deep-reasoning analysis — the most thorough model reads every window of the video. Better on fast cuts and on scenes with many people, and it does not invent cuts. Billed per duration bucket.",
|
|
4222
|
+
// Not `most-accurate`: that tag is Smart's (hand-counted edit list, always
|
|
4223
|
+
// refined). Ultra's strengths are the footage that trips the other tiers.
|
|
4224
|
+
useCases: ["video-analysis", "shot-list", "premium", "fast-cuts", "many-people", "no-invented-cuts"],
|
|
4225
|
+
pricing: [
|
|
4226
|
+
{ identifier: "video-analysis:gemini-4-carbon", credits: 1745, note: "10-min ceiling (no duration given)" },
|
|
4227
|
+
{ identifier: "video-analysis:gemini-4-carbon:60s", credits: 349 },
|
|
4228
|
+
{ identifier: "video-analysis:gemini-4-carbon:180s", credits: 375 },
|
|
4229
|
+
{ identifier: "video-analysis:gemini-4-carbon:360s", credits: 1054 },
|
|
4230
|
+
{ identifier: "video-analysis:gemini-4-carbon:600s", credits: 1745, note: "10-min ceiling" },
|
|
4231
|
+
],
|
|
4232
|
+
},
|
|
3900
4233
|
// Both mixed tiers are variants of the same advanced multi-engine analysis
|
|
3901
4234
|
// and share this ONE credit family (videoAnalysisCreditSegment maps both;
|
|
3902
4235
|
// plan internals live in the private analysis plugin).
|
|
@@ -3911,11 +4244,11 @@ const VIDEO_MODELS: Record<string, ModelCatalogEntry> = {
|
|
|
3911
4244
|
description: "Our most advanced analysis tier — multiple analysis engines combined into one result for maximum completeness and accuracy. Billed per duration bucket.",
|
|
3912
4245
|
useCases: ["video-analysis", "shot-list", "premium", "most-complete"],
|
|
3913
4246
|
pricing: [
|
|
3914
|
-
{ identifier: "video-analysis:mixed", credits:
|
|
3915
|
-
{ identifier: "video-analysis:mixed:60s", credits:
|
|
3916
|
-
{ identifier: "video-analysis:mixed:180s", credits:
|
|
3917
|
-
{ identifier: "video-analysis:mixed:360s", credits:
|
|
3918
|
-
{ identifier: "video-analysis:mixed:600s", credits:
|
|
4247
|
+
{ identifier: "video-analysis:mixed", credits: 1209, note: "10-min ceiling (no duration given)" },
|
|
4248
|
+
{ identifier: "video-analysis:mixed:60s", credits: 275 },
|
|
4249
|
+
{ identifier: "video-analysis:mixed:180s", credits: 298 },
|
|
4250
|
+
{ identifier: "video-analysis:mixed:360s", credits: 748 },
|
|
4251
|
+
{ identifier: "video-analysis:mixed:600s", credits: 1209, note: "10-min ceiling" },
|
|
3919
4252
|
],
|
|
3920
4253
|
},
|
|
3921
4254
|
// SMART — the accuracy tier. Since the 2026-08-03 hybrid re-plan it is a
|
|
@@ -3934,11 +4267,11 @@ const VIDEO_MODELS: Record<string, ModelCatalogEntry> = {
|
|
|
3934
4267
|
description: "Highest-accuracy analysis — a hybrid pass that blends a native skeleton read with several donor analysis rolls, then always refines the merged result to find shot boundaries and identify the cast by appearance. Best choice when the shot list will drive regeneration. Billed per duration bucket.",
|
|
3935
4268
|
useCases: ["video-analysis", "shot-list", "premium", "most-accurate"],
|
|
3936
4269
|
pricing: [
|
|
3937
|
-
{ identifier: "video-analysis:smart", credits:
|
|
3938
|
-
{ identifier: "video-analysis:smart:60s", credits:
|
|
3939
|
-
{ identifier: "video-analysis:smart:180s", credits:
|
|
3940
|
-
{ identifier: "video-analysis:smart:360s", credits:
|
|
3941
|
-
{ identifier: "video-analysis:smart:600s", credits:
|
|
4270
|
+
{ identifier: "video-analysis:smart", credits: 2097, note: "10-min ceiling (no duration given)" },
|
|
4271
|
+
{ identifier: "video-analysis:smart:60s", credits: 416 },
|
|
4272
|
+
{ identifier: "video-analysis:smart:180s", credits: 508 },
|
|
4273
|
+
{ identifier: "video-analysis:smart:360s", credits: 1280 },
|
|
4274
|
+
{ identifier: "video-analysis:smart:600s", credits: 2097, note: "10-min ceiling" },
|
|
3942
4275
|
],
|
|
3943
4276
|
},
|
|
3944
4277
|
// ── AI Audit (`video-audit`) ── a SEPARATE node from video-analysis (own
|
|
@@ -3983,11 +4316,11 @@ const VIDEO_MODELS: Record<string, ModelCatalogEntry> = {
|
|
|
3983
4316
|
description: "Same video-verified audit as AI Audit, but auto-runs a fast analysis first when none is wired in, then applies corrections under guards and returns a disclosed report. Billed per duration bucket.",
|
|
3984
4317
|
useCases: ["video-audit", "qa", "shot-list", "verification", "auto-analysis"],
|
|
3985
4318
|
pricing: [
|
|
3986
|
-
{ identifier: "video-audit:auto", credits:
|
|
3987
|
-
{ identifier: "video-audit:auto:60s", credits:
|
|
3988
|
-
{ identifier: "video-audit:auto:180s", credits:
|
|
3989
|
-
{ identifier: "video-audit:auto:360s", credits:
|
|
3990
|
-
{ identifier: "video-audit:auto:600s", credits:
|
|
4319
|
+
{ identifier: "video-audit:auto", credits: 1953, note: "10-min ceiling (no duration given)" },
|
|
4320
|
+
{ identifier: "video-audit:auto:60s", credits: 401 },
|
|
4321
|
+
{ identifier: "video-audit:auto:180s", credits: 483 },
|
|
4322
|
+
{ identifier: "video-audit:auto:360s", credits: 1197 },
|
|
4323
|
+
{ identifier: "video-audit:auto:600s", credits: 1953, note: "10-min ceiling" },
|
|
3991
4324
|
],
|
|
3992
4325
|
},
|
|
3993
4326
|
}
|
|
@@ -4031,6 +4364,65 @@ const DIALOGUE_V4_MAX_CHARS = 10000
|
|
|
4031
4364
|
*/
|
|
4032
4365
|
const V4_TURBO_MAX_CHARS = 10000
|
|
4033
4366
|
|
|
4367
|
+
/**
|
|
4368
|
+
* Gemini 3.8 (Flash-Lite, Flash) speech, one speaker: its own 70 voices, its
|
|
4369
|
+
* closed list of `<angle>` tone tags, a free-text style per line. No language
|
|
4370
|
+
* field, no timings, no neighbour text, no ElevenLabs levers. From the
|
|
4371
|
+
* provider's schema (2026-11-12); whether each tag is performed rather than
|
|
4372
|
+
* read aloud is to be checked on staging.
|
|
4373
|
+
*/
|
|
4374
|
+
const GEMINI_3_8_TTS_SHEET: TtsCapabilities = {
|
|
4375
|
+
audioTags: true,
|
|
4376
|
+
tagSyntax: "angle",
|
|
4377
|
+
tags: GEMINI_3_8_TONE_TAGS,
|
|
4378
|
+
ssmlBreaks: false,
|
|
4379
|
+
levers: [],
|
|
4380
|
+
languageCode: false,
|
|
4381
|
+
timestamps: false,
|
|
4382
|
+
stitching: false,
|
|
4383
|
+
maxChars: GEMINI_TTS_MAX_CHARS,
|
|
4384
|
+
// No picker: the language follows the text.
|
|
4385
|
+
languages: [],
|
|
4386
|
+
voices: GEMINI_3_8_TTS_VOICES,
|
|
4387
|
+
defaultVoice: GEMINI_3_8_DEFAULT_VOICE,
|
|
4388
|
+
direction: { styleMaxChars: TTS_DIRECTION_MAX_CHARS },
|
|
4389
|
+
ratePer1kChars: true,
|
|
4390
|
+
// Two-speaker dialogue (decided 2026-11-12): one or two voices, a style per line, filler words on request.
|
|
4391
|
+
dialogueSpeakers: 2,
|
|
4392
|
+
fillerWords: true,
|
|
4393
|
+
// A dialogue's file is chosen as one speaker's is: MP3 or the WAV as rendered (decided 2026-11-13).
|
|
4394
|
+
dialogueAnyOutputFormat: true,
|
|
4395
|
+
}
|
|
4396
|
+
|
|
4397
|
+
/**
|
|
4398
|
+
* Gemini 3.1 Flash and 2.5 Pro speech, one speaker: 30 voices, `[square]`
|
|
4399
|
+
* tone tags written freely (sent as written), and the speaker's style, pace,
|
|
4400
|
+
* accent and audio profile. Otherwise as the 3.8 sheet.
|
|
4401
|
+
*/
|
|
4402
|
+
const GEMINI_3_1_TTS_SHEET: TtsCapabilities = {
|
|
4403
|
+
audioTags: true,
|
|
4404
|
+
ssmlBreaks: false,
|
|
4405
|
+
levers: [],
|
|
4406
|
+
languageCode: false,
|
|
4407
|
+
timestamps: false,
|
|
4408
|
+
stitching: false,
|
|
4409
|
+
maxChars: GEMINI_TTS_MAX_CHARS,
|
|
4410
|
+
languages: [],
|
|
4411
|
+
voices: GEMINI_3_1_TTS_VOICES,
|
|
4412
|
+
defaultVoice: GEMINI_3_1_DEFAULT_VOICE,
|
|
4413
|
+
direction: {
|
|
4414
|
+
styles: GEMINI_3_1_TTS_STYLES,
|
|
4415
|
+
paces: GEMINI_3_1_TTS_PACES,
|
|
4416
|
+
accents: GEMINI_3_1_TTS_ACCENTS,
|
|
4417
|
+
audioProfileMaxChars: TTS_DIRECTION_MAX_CHARS,
|
|
4418
|
+
},
|
|
4419
|
+
ratePer1kChars: true,
|
|
4420
|
+
// Two-speaker dialogue (decided 2026-11-12): one or two voices, the style, pace, accent and profile per speaker.
|
|
4421
|
+
dialogueSpeakers: 2,
|
|
4422
|
+
// A dialogue's file is chosen as one speaker's is: MP3 or the WAV as rendered (decided 2026-11-13).
|
|
4423
|
+
dialogueAnyOutputFormat: true,
|
|
4424
|
+
}
|
|
4425
|
+
|
|
4034
4426
|
const AUDIO_MODELS: Record<string, ModelCatalogEntry> = {
|
|
4035
4427
|
// ── ElevenLabs TTS ──
|
|
4036
4428
|
"elevenlabs-v3": {
|
|
@@ -4179,6 +4571,86 @@ const AUDIO_MODELS: Record<string, ModelCatalogEntry> = {
|
|
|
4179
4571
|
languages: TTS_LANGS_MULTILINGUAL_V2,
|
|
4180
4572
|
},
|
|
4181
4573
|
},
|
|
4574
|
+
|
|
4575
|
+
// ── Google Gemini TTS (through KIE) ──
|
|
4576
|
+
// One id per model for both lanes: the same KIE model takes one speaker
|
|
4577
|
+
// (Text to Speech) and two (Text to Dialogue), and dialogue costs the same per
|
|
4578
|
+
// character as one speaker (decided 2026-11-12) — so `tts` AND `dialogue`,
|
|
4579
|
+
// one sheet and one price row, the same rate by construction.
|
|
4580
|
+
// Named voices of their own (never ElevenLabs'), no language field (the
|
|
4581
|
+
// language follows the text), no timings, no neighbour text. Priced per
|
|
4582
|
+
// 1,000 characters on the same count as every speech model (decided
|
|
4583
|
+
// 2026-11-12, provisional until measured on staging); the flat row is the
|
|
4584
|
+
// price of 1,000 characters.
|
|
4585
|
+
"gemini-3-8-flash-lite-tts": {
|
|
4586
|
+
id: "gemini-3-8-flash-lite-tts",
|
|
4587
|
+
kind: "audio",
|
|
4588
|
+
modes: ["tts", "dialogue"] as const,
|
|
4589
|
+
family: "Google",
|
|
4590
|
+
maker: "google",
|
|
4591
|
+
label: "Gemini 3.8 Flash-Lite TTS",
|
|
4592
|
+
series: "Gemini",
|
|
4593
|
+
description: "Google's lightest Gemini speech model and the cheapest speech on the platform — 70 named voices grouped by use, its own tone tags (laugh, sigh, short pause and more) and a free-text style direction. The language follows the text; up to 10,000 characters per request. Also voices a dialogue of two speakers.",
|
|
4594
|
+
useCases: ["tts", "voice-over", "narration", "fast", "budget"],
|
|
4595
|
+
features: ["audio-tags"],
|
|
4596
|
+
pricing: [
|
|
4597
|
+
{ identifier: "gemini-3-8-flash-lite-tts", credits: 4 },
|
|
4598
|
+
{ identifier: "gemini-3-8-flash-lite-tts:per-1k-chars", credits: 4, note: SPEECH_PER_1K_PRICE_NOTE },
|
|
4599
|
+
],
|
|
4600
|
+
tts: GEMINI_3_8_TTS_SHEET,
|
|
4601
|
+
},
|
|
4602
|
+
"gemini-3-8-flash-tts": {
|
|
4603
|
+
id: "gemini-3-8-flash-tts",
|
|
4604
|
+
kind: "audio",
|
|
4605
|
+
modes: ["tts", "dialogue"] as const,
|
|
4606
|
+
family: "Google",
|
|
4607
|
+
maker: "google",
|
|
4608
|
+
label: "Gemini 3.8 Flash TTS",
|
|
4609
|
+
series: "Gemini",
|
|
4610
|
+
description: "Google's Gemini 3.8 speech model — the same 70 voices, tone tags and free-text style direction as Flash-Lite, from the larger model of the generation. The language follows the text; up to 10,000 characters per request. Also voices a dialogue of two speakers.",
|
|
4611
|
+
useCases: ["tts", "voice-over", "narration", "expressive"],
|
|
4612
|
+
features: ["audio-tags"],
|
|
4613
|
+
pricing: [
|
|
4614
|
+
{ identifier: "gemini-3-8-flash-tts", credits: 6 },
|
|
4615
|
+
{ identifier: "gemini-3-8-flash-tts:per-1k-chars", credits: 6, note: SPEECH_PER_1K_PRICE_NOTE },
|
|
4616
|
+
],
|
|
4617
|
+
// A one-off clone (`replica`, decided 2026-11-13) on 3.8 Flash alone: its Flash-Lite sibling refuses it.
|
|
4618
|
+
tts: { ...GEMINI_3_8_TTS_SHEET, replica: true },
|
|
4619
|
+
},
|
|
4620
|
+
"gemini-3-1-flash-tts": {
|
|
4621
|
+
id: "gemini-3-1-flash-tts",
|
|
4622
|
+
kind: "audio",
|
|
4623
|
+
modes: ["tts", "dialogue"] as const,
|
|
4624
|
+
family: "Google",
|
|
4625
|
+
maker: "google",
|
|
4626
|
+
label: "Gemini 3.1 Flash TTS",
|
|
4627
|
+
series: "Gemini",
|
|
4628
|
+
description: "Google's Gemini 3.1 speech model — 30 named voices, [audio tags] written freely, and a speaking style, pace, accent and audio profile for the speaker. The language follows the text; up to 10,000 characters per request. Also voices a dialogue of two speakers.",
|
|
4629
|
+
useCases: ["tts", "voice-over", "narration", "expressive"],
|
|
4630
|
+
features: ["audio-tags"],
|
|
4631
|
+
pricing: [
|
|
4632
|
+
{ identifier: "gemini-3-1-flash-tts", credits: 13 },
|
|
4633
|
+
{ identifier: "gemini-3-1-flash-tts:per-1k-chars", credits: 13, note: SPEECH_PER_1K_PRICE_NOTE },
|
|
4634
|
+
],
|
|
4635
|
+
tts: GEMINI_3_1_TTS_SHEET,
|
|
4636
|
+
},
|
|
4637
|
+
"gemini-2-5-pro-tts": {
|
|
4638
|
+
id: "gemini-2-5-pro-tts",
|
|
4639
|
+
kind: "audio",
|
|
4640
|
+
modes: ["tts", "dialogue"] as const,
|
|
4641
|
+
family: "Google",
|
|
4642
|
+
maker: "google",
|
|
4643
|
+
label: "Gemini 2.5 Pro TTS",
|
|
4644
|
+
series: "Gemini",
|
|
4645
|
+
description: "Google's Gemini 2.5 Pro speech model — the same 30 voices, [audio tags] and speaker controls (style, pace, accent, audio profile) as Gemini 3.1 Flash TTS, from the Pro model. The language follows the text; up to 10,000 characters per request. Also voices a dialogue of two speakers.",
|
|
4646
|
+
useCases: ["tts", "voice-over", "narration", "expressive", "long-form"],
|
|
4647
|
+
features: ["audio-tags"],
|
|
4648
|
+
pricing: [
|
|
4649
|
+
{ identifier: "gemini-2-5-pro-tts", credits: 16 },
|
|
4650
|
+
{ identifier: "gemini-2-5-pro-tts:per-1k-chars", credits: 16, note: SPEECH_PER_1K_PRICE_NOTE },
|
|
4651
|
+
],
|
|
4652
|
+
tts: GEMINI_3_1_TTS_SHEET,
|
|
4653
|
+
},
|
|
4182
4654
|
"elevenlabs-dialogue": {
|
|
4183
4655
|
id: "elevenlabs-dialogue",
|
|
4184
4656
|
kind: "audio",
|
|
@@ -4415,6 +4887,48 @@ const AUDIO_MODELS: Record<string, ModelCatalogEntry> = {
|
|
|
4415
4887
|
useCases: ["music", "song", "vocals"],
|
|
4416
4888
|
pricing: [{ identifier: "suno", credits: 30, note: "per generation" }],
|
|
4417
4889
|
},
|
|
4890
|
+
// Suno v4.5, v4.5 All and v4.5+ (decided 2026-11-12): listed so a
|
|
4891
|
+
// deployment's model restriction can name them, each with a price row of its
|
|
4892
|
+
// own (SUNO_VERSION_CREDIT_KEYS) at the price those runs were charged before.
|
|
4893
|
+
"suno-v4_5": {
|
|
4894
|
+
id: "suno-v4_5",
|
|
4895
|
+
kind: "audio",
|
|
4896
|
+
modes: ["music"] as const,
|
|
4897
|
+
family: "Suno",
|
|
4898
|
+
maker: "suno",
|
|
4899
|
+
label: "Suno v4.5",
|
|
4900
|
+
series: "Suno",
|
|
4901
|
+
dataValue: "V4_5",
|
|
4902
|
+
description: "Suno v4.5 — improved quality and coherence over v4.",
|
|
4903
|
+
useCases: ["music", "song", "vocals"],
|
|
4904
|
+
pricing: [{ identifier: "suno-v4_5", credits: 30, note: "per generation" }],
|
|
4905
|
+
},
|
|
4906
|
+
"suno-v4_5all": {
|
|
4907
|
+
id: "suno-v4_5all",
|
|
4908
|
+
kind: "audio",
|
|
4909
|
+
modes: ["music"] as const,
|
|
4910
|
+
family: "Suno",
|
|
4911
|
+
maker: "suno",
|
|
4912
|
+
label: "Suno v4.5 All",
|
|
4913
|
+
series: "Suno",
|
|
4914
|
+
dataValue: "V4_5ALL",
|
|
4915
|
+
description: "Suno v4.5 All — full instrument and vocal generation.",
|
|
4916
|
+
useCases: ["music", "song", "vocals"],
|
|
4917
|
+
pricing: [{ identifier: "suno-v4_5all", credits: 30, note: "per generation" }],
|
|
4918
|
+
},
|
|
4919
|
+
"suno-v4_5plus": {
|
|
4920
|
+
id: "suno-v4_5plus",
|
|
4921
|
+
kind: "audio",
|
|
4922
|
+
modes: ["music"] as const,
|
|
4923
|
+
family: "Suno",
|
|
4924
|
+
maker: "suno",
|
|
4925
|
+
label: "Suno v4.5 Plus",
|
|
4926
|
+
series: "Suno",
|
|
4927
|
+
dataValue: "V4_5PLUS",
|
|
4928
|
+
description: "Suno v4.5 Plus — richer production than v4.5.",
|
|
4929
|
+
useCases: ["music", "song", "vocals"],
|
|
4930
|
+
pricing: [{ identifier: "suno-v4_5plus", credits: 30, note: "per generation" }],
|
|
4931
|
+
},
|
|
4418
4932
|
"suno-v5": {
|
|
4419
4933
|
id: "suno-v5",
|
|
4420
4934
|
kind: "audio",
|