Fluxion API

API HTTP untuk generate gambar, video, dan audio (text-to-speech) secara terprogram. Autentikasi pakai API key — minta admin workspace buat bikinin key di halaman API Keys, lalu kirim sebagai header Authorization: Bearer <key> di setiap request.

POST/api/generate

Generate gambar dari teks. Respons: { pathname, prompt, style, steps, createdAt, createdBy, model, ratio }. Unduh file hasilnya di /images/{pathname} — endpoint ini juga butuh header Authorization, bukan URL publik.

Field	Tipe	Keterangan
prompt	string	wajib. Boleh Bahasa Indonesia — otomatis diterjemahkan.
model	string	lihat tabel model di bawah. Default: model pertama.
steps	number	custom, dibatasi min/max tiap model
style	enum	none, photorealistic, comic-book, neon-punk, isometric, line-art, pixel-art, 3d-model
ratio	enum	1:1, 16:9, 9:16 — default 1:1

Model tersedia

id (model)	Steps (min/default/max)
cf/@cf/black-forest-labs/flux-1-schnell	4 / 4 / 8
cf/@cf/black-forest-labs/flux-2-dev	4 / 8 / 20
cf/@cf/black-forest-labs/flux-2-klein-9b	4 / 4 / 8
cf/@cf/black-forest-labs/flux-2-klein-4b	4 / 4 / 8
cf/@cf/leonardo/lucid-origin	4 / 6 / 12
cf/@cf/leonardo/phoenix-1.0	4 / 6 / 12
cf/@cf/bytedance/stable-diffusion-xl-lightning	2 / 4 / 8
qwen/qwen-image-2.0-pro	4 / 4 / 4
qwen/qwen-image-2.0	4 / 4 / 4
qwen/qwen-image-max	4 / 4 / 4
qwen/qwen-image-plus	4 / 4 / 4
qwen/z-image-turbo	4 / 4 / 4
qwen/wan2.7-image-pro	4 / 4 / 4
qwen/wan2.7-image	4 / 4 / 4
qwen/qwen-image-3.0-pro	4 / 4 / 4
qwen/qwen-image-3.0	4 / 4 / 4
qwen/wan2.6-t2i	4 / 4 / 4
gs/nano-banana-pro	4 / 4 / 4
gs/nano-banana-2	4 / 4 / 4
gs/nano-banana-2-flash-lite	4 / 4 / 4
gs/gpt-image-2	4 / 4 / 4
gs/fal-ai/bytedance/seedream/v5/pro	4 / 4 / 4
gs/fal-ai/bytedance/seedream/v5/lite	4 / 4 / 4
gs/xai/grok-imagine-image	4 / 4 / 4
gs/qwen-image-3	4 / 4 / 4
gs/fal-ai/flux-2-pro	4 / 4 / 4
gs/fal-ai/z-image/turbo	4 / 4 / 4

curl -X POST https://domain-kamu.com/api/generate \
  -H "Authorization: Bearer sk-fluxion-..." \
  -H "Content-Type: application/json" \
  -d '{"prompt":"a red fox in a forest","model":"cf/@cf/black-forest-labs/flux-1-schnell","steps":4,"ratio":"1:1","style":"photorealistic"}'

POST/api/video/generate

Generate video dari teks (text-to-video). Bersifat async — request ini cuma submit job.

Field	Tipe	Keterangan
prompt	string	wajib
model	string	lihat tabel model di bawah
ratio	enum	1:1, 16:9, 9:16 — default 16:9
duration	number	2–15 detik, default 5 (diabaikan pada model yang gak support durasi custom)

Respons awal: { taskId, status: "PENDING" }. Poll GET /api/video/status/{taskId} tiap beberapa detik sampai status jadi SUCCEEDED (lalu ambil field pathname) atau FAILED.

Model tersedia

wan2.6-t2v	Wan 2.6 T2V
wan2.7-t2v	Wan 2.7 T2V
wan2.5-t2v-preview	Wan 2.5 T2V (audio)
wan2.2-t2v-plus	Wan 2.2 T2V Plus
wan2.1-t2v-turbo	Wan 2.1 T2V Turbo
wan2.7-t2v-2026-06-12	Wan 2.7 T2V (06-12)
wan2.7-t2v-2026-04-25	Wan 2.7 T2V (04-25)
happyhorse-1.1-t2v	HappyHorse 1.1 T2V

curl -X POST https://domain-kamu.com/api/video/generate \
  -H "Authorization: Bearer sk-fluxion-..." \
  -H "Content-Type: application/json" \
  -d '{"prompt":"a cat playing with yarn","model":"wan2.7-t2v","ratio":"16:9","duration":5}'

curl https://domain-kamu.com/api/video/status/{taskId} \
  -H "Authorization: Bearer sk-fluxion-..."

POST/api/image-to-video/generate

Gerakkan gambar diam jadi video (image-to-video). Async, sama seperti /api/video/generate: poll GET /api/video/status/{taskId} sampai selesai.

Field	Tipe	Keterangan
prompt	string	wajib — deskripsi gerakan yang diinginkan
image	string	wajib — base64 data URL, contoh: `data:image/jpeg;base64,...`
model	string	lihat tabel model di bawah
duration	number	2–15 detik, default 5

Model tersedia

wan2.7-i2v	Wan 2.7 I2V
wan2.6-i2v	Wan 2.6 I2V
wan2.6-i2v-flash	Wan 2.6 I2V Flash
wan2.5-i2v-preview	Wan 2.5 I2V (audio)
wan2.2-i2v-flash	Wan 2.2 I2V Flash
wan2.1-i2v-turbo	Wan 2.1 I2V Turbo
wan2.7-i2v-2026-04-25	Wan 2.7 I2V (04-25)

curl -X POST https://domain-kamu.com/api/image-to-video/generate \
  -H "Authorization: Bearer sk-fluxion-..." \
  -H "Content-Type: application/json" \
  -d '{"prompt":"hair blowing in the wind, slow zoom in","model":"wan2.7-i2v","duration":5,"image":"data:image/jpeg;base64,/9j/4AAQ..."}'

curl https://domain-kamu.com/api/video/status/{taskId} \
  -H "Authorization: Bearer sk-fluxion-..."

POST/api/edit-image/generate

Edit gambar yang sudah ada berdasarkan instruksi teks (ganti warna, tambah/hapus objek, ubah suasana, dll). Sinkron — respons langsung berisi gambar hasil, gak perlu polling.

Field	Tipe	Keterangan
prompt	string	wajib — instruksi perubahan
image	string	wajib — base64 data URL
model	string	lihat tabel model di bawah

Model tersedia

qwen-image-edit-plus-2025-12-15	Qwen Image Edit Plus
qwen-image-edit-max	Qwen Image Edit Max
qwen-image-edit-max-2026-01-16	Qwen Image Edit Max (01-16)
qwen-image-edit-plus	Qwen Image Edit Plus (latest)
qwen-image-edit-plus-2025-10-30	Qwen Image Edit Plus (10-30)
qwen-image-edit	Qwen Image Edit
gs/nano-banana-pro	Nano Banana Pro (4K)
gs/nano-banana-2	Nano Banana 2
gs/nano-banana-2-flash-lite	Nano Banana 2 Flash Lite
gs/gpt-image-2	GPT Image 2
gs/fal-ai/bytedance/seedream/v5/pro	Seedream v5 Pro
gs/fal-ai/bytedance/seedream/v5/lite	Seedream v5 Lite
gs/xai/grok-imagine-image	Grok Imagine
gs/qwen-image-3	Qwen Image 3 (Genspark)
gs/fal-ai/flux-2-pro	Flux 2 Pro
gs/fal-bria-rmbg	Hapus Background
gs/fal-ai/recraft-clarity-upscale	Upscale
gs/fal-ai/image-editing/text-removal	Hapus Teks / Watermark
gs/flux-pro/outpaint	Perluas Gambar (Outpaint)

curl -X POST https://domain-kamu.com/api/edit-image/generate \
  -H "Authorization: Bearer sk-fluxion-..." \
  -H "Content-Type: application/json" \
  -d '{"prompt":"change the shirt color to blue","model":"qwen-image-edit-plus-2025-12-15","image":"data:image/jpeg;base64,/9j/4AAQ..."}'

POST/api/reference-video/generate

Buat video pakai 1-9 gambar referensi (karakter, gaya, objek) — hasil video konsisten dengan referensinya. Async, sama seperti /api/video/generate: poll GET /api/video/status/{taskId} sampai selesai.

Field	Tipe	Keterangan
prompt	string	wajib
images	string[]	wajib — 1-9 base64 data URL
model	string	lihat tabel model di bawah
duration	number	2–15 detik, default 5

Model tersedia

wan2.7-r2v-2026-06-12	Wan 2.7 R2V
happyhorse-1.1-r2v	HappyHorse 1.1 R2V

curl -X POST https://domain-kamu.com/api/reference-video/generate \
  -H "Authorization: Bearer sk-fluxion-..." \
  -H "Content-Type: application/json" \
  -d '{"prompt":"the character walks through a sunny park","model":"wan2.7-r2v-2026-06-12","duration":5,"images":["data:image/jpeg;base64,/9j/4AAQ..."]}'

curl https://domain-kamu.com/api/video/status/{taskId} \
  -H "Authorization: Bearer sk-fluxion-..."

POST/api/video-edit/generate

Edit video yang sudah ada pakai instruksi teks (ganti elemen, tambah efek, ubah suasana). Async, sama seperti /api/video/generate: poll GET /api/video/status/{taskId} sampai selesai.

Field	Tipe	Keterangan
prompt	string	wajib — instruksi perubahan
video	string	wajib — base64 data URL video, contoh: `data:video/mp4;base64,...`
model	string	lihat tabel model di bawah

Model tersedia

wan2.7-videoedit	Wan 2.7 Video Edit
happyhorse-1.0-video-edit	HappyHorse 1.0 Video Edit

curl -X POST https://domain-kamu.com/api/video-edit/generate \
  -H "Authorization: Bearer sk-fluxion-..." \
  -H "Content-Type: application/json" \
  -d '{"prompt":"make it snow","model":"wan2.7-videoedit","video":"data:video/mp4;base64,AAAAIGZ0eXA..."}'

curl https://domain-kamu.com/api/video/status/{taskId} \
  -H "Authorization: Bearer sk-fluxion-..."

POST/api/audio/generate

Text-to-speech, support Bahasa Indonesia native (16 bahasa). Respons: { pathname, text, voice, createdAt, createdBy }.

Field	Tipe	Keterangan
text	string	wajib, max 2000 karakter
voice	string	lihat tabel suara di bawah

Suara tersedia

longanlingxin	Lingxin — Perempuan, hangat
longanlufeng	Lufeng — Laki-laki, ceria
longanfengyue	Fengyue — Perempuan, natural
longanyuanfei	Yuanfei — Perempuan, anggun
longanlingxi	Lingxi — Perempuan, imut
longanxiaoxin	Xiaoxin — Perempuan, ceria
longanhuan_v3.6	Huan — Perempuan
longchuanshu_v3.6	Chuanshu — Laki-laki, dewasa
loongmary	Mary — Perempuan, aksen Inggris
loongeva_v3.6	Eva — Perempuan, aksen Amerika
loongjohn	John — Laki-laki, aksen Amerika
longjielidou_v3.6	Jielidou — Anak laki-laki
longpaopao_v3.6	Paopao — Anak perempuan
longhuohuo_v3.6	Huohuo — Anak laki-laki, jahil

curl -X POST https://domain-kamu.com/api/audio/generate \
  -H "Authorization: Bearer sk-fluxion-..." \
  -H "Content-Type: application/json" \
  -d '{"text":"Selamat pagi, apa kabar?","voice":"longanlingxin"}'

Endpoint baca (GET)

/api/gallery	riwayat gambar tim
/api/videos	riwayat video yang sudah selesai (text-to-video maupun image-to-video)
/api/audios	riwayat audio
/api/models	daftar model gambar (sama seperti tabel di atas, plus URL preview)
/api/video-models	daftar model text-to-video
/api/image-to-video-models	daftar model image-to-video
/api/voices	daftar suara TTS
/api/styles	daftar gaya gambar (none, photorealistic, comic-book, dst)
/images/{pathname}	ambil file hasil generate (gambar/video/audio) — bukan URL publik, tetap butuh header `Authorization`

POST/api/improve-prompt

Perbaiki/perluas prompt sebelum generate. Perilakunya beda tergantung type: image menghasilkan "mega prompt" terstruktur (composition, environment, style, mood, camera, negative prompt), video/image-to-video menghasilkan breakdown scene-by-scene, audio cuma merapikan tata bahasa/tanda baca tanpa mengubah makna atau menerjemahkan.

Field	Tipe	Keterangan
prompt	string	wajib
type	enum	image, video, image-to-video, audio — default image

curl -X POST https://domain-kamu.com/api/improve-prompt \
  -H "Authorization: Bearer sk-fluxion-..." \
  -H "Content-Type: application/json" \
  -d '{"prompt":"kucing lari di taman","type":"image"}'

Autentikasi & error

401	API key hilang/salah — header `Authorization: Bearer <key>` wajib di semua request
403	API key gak punya akses ke endpoint admin