Setup
import { createVarg } from "vargai/ai"
const varg = createVarg({ apiKey: process.env.VARG_API_KEY! })
varg client. 1 credit = 1 cent.
For detailed documentation, parameters, and examples for each model, see the Models section.
Quick Reference
| Category | Recommended | Code | Credits | Docs |
|---|---|---|---|---|
| Image (versatile) | Nano Banana Pro | varg.imageModel("nano-banana-pro") | 5 | Details |
| Image (quality) | Flux Pro | varg.imageModel("flux-pro") | 10 | Details |
| Image (editing) | Nano Banana Edit | varg.imageModel("nano-banana-pro/edit") | 5 | Details |
| Image (characters) | Soul | varg.imageModel("soul") | 15 | Details |
| Image (cheapest) | Grok Imagine | varg.imageModel("grok-imagine-image") | 3 | Details |
| Video (best) | Kling V3 | varg.videoModel("kling-v3") | 150 | Details |
| Video (premium) | Seedance 2 | varg.videoModel("seedance_2") | 160 (5s@720p) | Details |
| Video (budget) | Wan 2.5 | varg.videoModel("wan-2.5") | 80 | Details |
| Video (cheapest) | LTX | varg.videoModel("ltx-2-19b-distilled") | 50 | Details |
| Lipsync (simple) | VEED Fabric | varg.videoModel("veed-fabric-1.0") | 100 | Details |
| Lipsync (best) | Sync V2 Pro | varg.videoModel("sync-v2-pro") | 80 | Details |
| Speech (best) | Eleven V3 | varg.speechModel("eleven_v3") | 25 | Details |
| Speech (multi-lang) | Eleven Multilingual V2 | varg.speechModel("eleven_multilingual_v2") | 25 | Details |
| Music | Music V1 | varg.musicModel() | 30 | Details |
| Transcription | Groq Whisper Turbo | varg.transcriptionModel("groq-whisper-large-v3-turbo") | 3 | Details |
| Image upscale | SeedVR | varg.imageModel("seedvr") | 10 | Details |
| Video upscale | Sima | varg.videoModel("sima-video-upscaler") | 15 | Details |
All models by category
Video Models
| Model | Code | Duration | Credits | Docs |
|---|---|---|---|---|
| Seedance 2 | varg.videoModel("seedance_2") | 4-15s or auto | per-second, 160 (5s@720p) | Details |
| Seedance 2 Mini | varg.videoModel("seedance_2_mini") | 4-15s (int) | per-second, 74 (5s@720p) | Details |
| Seedance 2 Less Restriction | varg.videoModel("seedance_2_less_restriction") | 4-15s (int) | per-second, 116 (5s@720p) — permissive review, explicit opt-in | Details |
| Kling V3 Pro | varg.videoModel("kling-v3") | 3-15s (int) | 150 | Details |
| Kling V3 Standard | varg.videoModel("kling-v3-standard") | 3-15s (int) | 100 | Details |
| Sora 2 | varg.videoModel("sora-2") | varies | 100 | Details |
| Sora 2 Pro | varg.videoModel("sora-2-pro") | varies | 300 | Details |
| Sora 2 Remix | varg.videoModel("sora-2-remix") | varies | 100 | Details |
| Grok Imagine | varg.videoModel("grok-imagine") | 1-15s | 100 | Details |
| Grok Imagine Edit | varg.videoModel("grok-imagine-edit") | varies | 100 | Details |
| Wan 2.5 | varg.videoModel("wan-2.5") | 3-10s | 80 | Details |
| Minimax | varg.videoModel("minimax") | 5-10s | 80 | Details |
| Wan 2.5 Preview | varg.videoModel("wan-2.5-preview") | 3-10s | 60 | Details |
| LTX | varg.videoModel("ltx-2-19b-distilled") | varies | 50 | Details |
| HeyGen Avatar | varg.videoModel("heygen-avatar") | varies | 300 | — |
| HeyGen Avatar IV | varg.videoModel("heygen-avatar-iv") | varies | 300 | — |
Seedance 2 bills per second; estimates are exact when you pass a numeric
duration ("auto" reserves the worst case). The budget tier is seedance_2_mini (5s@480p = 37 credits, cheapest). Exact model_keys remain available for advanced route pinning — see Seedance 2 for the full route table.Image Models
| Model | Code | Credits | Docs |
|---|---|---|---|
| Nano Banana Pro | varg.imageModel("nano-banana-pro") | 5 | Details |
| Nano Banana Edit | varg.imageModel("nano-banana-pro/edit") | 5 | Details |
| Nano Banana 2 | varg.imageModel("nano-banana-2") | 5 | Details |
| Nano Banana 2 Edit | varg.imageModel("nano-banana-2/edit") | 5 | Details |
| Flux Pro | varg.imageModel("flux-pro") | 10 | Details |
| Flux Dev | varg.imageModel("flux-dev") | 8 | Details |
| Flux Schnell | varg.imageModel("flux-schnell") | 5 | Details |
| Phota | varg.imageModel("phota") | 9 | Details |
| Phota Edit | varg.imageModel("phota/edit") | 9 | Details |
| Phota Enhance | varg.imageModel("phota/enhance") | 13 | Details |
| Grok Imagine Image | varg.imageModel("grok-imagine-image") | 3 | Details |
| Grok Imagine Image Edit | varg.imageModel("grok-imagine-image/edit") | 3 | Details |
| Recraft V3 | varg.imageModel("recraft-v3") | 10 | Details |
| Recraft V4 Pro | varg.imageModel("recraft-v4-pro") | 30 | Details |
| Qwen Angles | varg.imageModel("qwen-angles") | 8 | Details |
| Qwen Image 2 | varg.imageModel("qwen-image-2") | 8 | Details |
| Reve | varg.imageModel("reve") | 5 | — |
| Seedream 5 Lite | varg.imageModel("seedream_5_lite") | 4 | Details |
| Seedream V4.5 Edit | varg.imageModel("seedream-v4.5/edit") | 10 | Details |
| Soul | varg.imageModel("soul") | 15 | Details |
Lipsync Models
| Model | Code | Credits | Docs |
|---|---|---|---|
| VEED Fabric 1.0 | varg.videoModel("veed-fabric-1.0") | 100 | Details |
| Sync V2 Pro | varg.videoModel("sync-v2-pro") | 80 | Details |
| Sync V2 | varg.videoModel("sync-v2") | 50 | Details |
| OmniHuman V1.5 | varg.videoModel("omnihuman-v1.5") | 100 | Details |
| Sync V3 | varg.videoModel("sync-v3") | 100 | Details |
| Lipsync | varg.videoModel("lipsync") | 50 | Details |
| LTX-2 A2V | varg.videoModel("ltx-2-a2v") | 50 | — |
Upscale Models
| Model | Code | Credits | Docs |
|---|---|---|---|
| SeedVR (image) | varg.imageModel("seedvr") | 10 | Details |
| Topaz (image) | varg.imageModel("topaz") | 15 | Details |
| Recraft Clarity (image) | varg.imageModel("recraft-clarity") | 8 | Details |
| Clarity Upscaler (image) | varg.imageModel("clarity-upscaler") | 8 | Details |
| CCSR (image) | varg.imageModel("ccsr") | 5 | Details |
| Aura SR (image) | varg.imageModel("aura-sr") | 5 | Details |
| SeedVR (video) | varg.videoModel("seedvr-video") | 30 | Details |
| Topaz (video) | varg.videoModel("topaz-video") | 50 | Details |
| ByteDance (video) | varg.videoModel("bytedance-upscaler") | 25 | Details |
| Sima (video) | varg.videoModel("sima-video-upscaler") | 15 | Details |
Speech Models
| Model | Code | Credits | Docs |
|---|---|---|---|
| Eleven V3 | varg.speechModel("eleven_v3") | 25 | Details |
| Eleven Multilingual V2 | varg.speechModel("eleven_multilingual_v2") | 25 | Details |
| Turbo | varg.speechModel("turbo") | 20 | Details |
| Eleven Turbo V2 | varg.speechModel("eleven_turbo_v2") | 20 | Details |
| Eleven Turbo V2.5 | varg.speechModel("eleven_turbo_v2_5") | 20 | Details |
| Eleven Flash V2 | varg.speechModel("eleven_flash_v2") | 20 | Details |
| Eleven Flash V2.5 | varg.speechModel("eleven_flash_v2_5") | 20 | Details |
Music
| Model | Code | Credits | Docs |
|---|---|---|---|
| Music V1 | varg.musicModel() | 30 | Details |
Transcription
| Model | Code | Credits | Docs |
|---|---|---|---|
| Whisper | varg.transcriptionModel("whisper") | 10 | Details |
| Whisper Large V3 | varg.transcriptionModel("whisper-large-v3") | 10 | Details |
| Groq Whisper | varg.transcriptionModel("groq-whisper") | 5 | Details |
| Groq Whisper Large V3 | varg.transcriptionModel("groq-whisper-large-v3") | 5 | Details |
| Groq Whisper Turbo | varg.transcriptionModel("groq-whisper-large-v3-turbo") | 3 | Details |
BYOK (Direct Provider Access)
If you have your own provider keys, you can use them directly:import { fal, elevenlabs, higgsfield, replicate } from "vargai/ai"
fal.imageModel("nano-banana-pro")
fal.videoModel("kling-v3")
elevenlabs.speechModel("eleven_v3")
elevenlabs.musicModel()
higgsfield.imageModel("soul")
.env. See BYOK for details.