跳至主要內容
Kyeo AI
AI 音訊模型
目前無法使用

Gemini 3.8 Flash-Lite TTS

Gemini 3.8 Flash-Lite TTS 目前無法用於新的生成任務；本頁提供模型狀態資訊。

目前暫不支援線上使用

此頁面僅供查看模型功能、適用情境與公開資料；該模型不會出現在首頁工作台，也無法提交生成任務。

查看可用模型

示例

Audio scene

FORMAT: 10 seconds, square 1:1, live action workplace comedy, grounded cinematic realism, hard cuts, rapid escalation. CAST ARI — male, 30, messy dark hair, nervous expressive eyes, pale blue striped pajamas, white tee, navy robe, grey slippers. MAYA — female, 42, sharp calm face, shoulder-length bob, camel trench coat, black suit, ivory blouse, burgundy heels, completely unshakable. DEV — male, 28, thin glasses, navy waistcoat, white shirt, green tie, clipboard, barely containing laughter. SET Compact apartment. Ari has tried to make it look like an ordinary messy home office. Sofa, laptop, tissues, coffee mug, blanket. Behind one curtain is secretly a professional fashion-shoot setup: ring light, softbox, backdrop, tripod, garment rack and several extravagant outfits. SHOT 1 — WIDE Front door opens. MAYA and DEV enter unexpectedly. ARI is sitting on the sofa in pajamas with his laptop open. Ari quickly closes it. MAYA: “Working from home?” ARI, innocent: “Very productive.” SHOT 2 — INSERT The closed laptop suddenly announces loudly: COMPUTER VOICE: “Next model to set, please.” Ari freezes. Dev slowly looks toward him. SHOT 3 — CLOSE-UP ARI reaches for the laptop. ARI: “YouTube.” A professional CAMERA FLASH explodes from behind the curtain. Everyone freezes. SHOT 4 — MEDIUM Maya calmly walks to the curtain. ARI: “Maya—” She pulls it open. Behind it: photographer, stylist and makeup artist standing frozen around a MODEL wearing a dramatic gold tuxedo. Full professional fashion set. SHOT 5 — RAPID HARD CUTS Same camera position. FLASH. Ari posing in a white faux-fur coat. CUT. FLASH. Ari posing in a hot-pink cowboy suit. CUT. FLASH. Ari posing in a metallic silver jacket and sunglasses. Each outfit is a completely separate practical shot. NO MORPHING. SHOT 6 — PRESENT TIME Back to Ari in pajamas. Silence. Maya looks at the studio. Then at Ari. MAYA: “Productive.” ARI nods seriously. ARI: “Very.” SHOT 7 — FINAL WIDE Dev notices Ari is still wearing glittering silver fashion boots underneath his pajama trousers. Dev points down. Ari slowly looks. Maya holds out her hand. MAYA: “Laptop.” Ari reluctantly gives it to her. The laptop screen wakes. CAMERA VOICE: “Ari, wardrobe four!” Ari closes his eyes. Huge CAMERA FLASH. CUT TO BLACK AT 10 SECONDS. AUDIO English dialogue with natural lip sync. Door opening, footsteps, laptop click, computer voice, curtain rings, camera shutters, flash recharge sound, fabric movement, one suppressed laugh from Dev and natural apartment room tone only. LOOK Soft daylight from camera left, realistic skin texture, subtle handheld micro-movement, shallow cinematic depth of field, 24fps natural motion blur, restrained performances, identical faces, wardrobe and apartment geography throughout. COMEDY STYLE Deadpan workplace confrontation. Maya never raises her voice or visibly reacts. Ari keeps inventing increasingly weak explanations while physical evidence destroys each lie. NEGATIVE No music, no score, no singing, no narration, no subtitles, no captions, no visible written text, no watermark, no duplicate characters, no extra limbs, no warped hands, no costume morphing, no magical transformations, no surreal transitions, no slow motion, no exaggerated cartoon reactions, no impossible camera movement.

Audio scene

Create a 15 second cinematic live action vertical salon fashion film in 9:16. Ethan is a 30 to 35 year old man with warm medium skin. Short dark brown hair. Brown eyes. Light stubble. Keep his face identical in every shot. Maya is a 28 to 32 year old stylist with warm light medium skin. Shoulder length dark chestnut hair. Brown eyes. Keep her identity identical throughout. Use a premium modern salon with warm lighting. Walnut walls. Black mirror. Burgundy leather chair. Maya asks “What are we becoming today?” Ethan replies “Surprise me.” Each salon tool creates a fast match cut into a new hairstyle and outfit. Spray bottle transforms Ethan into elegant black tie fashion. Comb transforms him into leather rebel style. Towel transforms him into 1970s luxury fashion. Hair dryer transforms him into modern street luxury. Styling cream transforms him into avant garde fashion. Scissors and comb create several rapid extra outfit changes. Ethan asks “How many versions do I get?” Maya replies “Until one feels like you.” Final look is natural and confident. Dark textured hair. Forest green suede overshirt. White shirt. Dark tailored trousers. Steel watch. Ethan says “That one.” Maya removes the cape and says “I know.” Use multiple close shots. Handheld camera. Mirror reflections. Foreground salon tools. Natural skin and fabric texture. Realistic movement. Fast physical transitions. No music. Only English dialogue and realistic salon sound effects. No subtitles. No logos. No text. No fantasy effects. No identity drift. No extra people. No deformed hands.

Audio scene

Create a 15 second photorealistic fashion performance video in 16:9, inspired by an intimate handheld phone recording but framed cinematically for widescreen. A young adult woman stands indoors in front of deep red curtains, a warm woven brown curtain panel, pale blue gray walls and a dark wooden doorway. She wears a white sleeveless textured blouse with delicate silver embellishment, a cream drape with rich red patterned borders wrapped like a sari, stacked metallic bangles, ornate silver drop earrings, a slim decorative waist chain and subtle arm jewelry. Her shoulder length black hair is softly layered with a natural center part. Makeup is polished but believable, with defined eyeliner, neutral lips and realistic skin texture. Begin in a stable medium close shot at chest height. She looks directly into the lens and softly lip syncs to the music with calm confidence. During the first three seconds she raises both hands toward her hair, brushes it back from her temples and briefly places her hands behind her head. Keep the motion imperfect and human, with tiny pauses, uneven wrist timing, jewelry inertia, loose fabric movement and small changes in breathing. From seconds three to seven she lowers her hands, lightly touches her hair and upper chest, shifts her weight from one hip to the other and makes gentle expressive gestures toward the camera. Her face moves through restrained longing, softness and a brief smile. The camera remains handheld with subtle micro sway and minor autofocus breathing, never robotic or perfectly stabilized. From seconds seven to ten she points lightly toward herself, then toward the viewer, steps a little closer and lets the framing tighten naturally from medium close to close up. Keep direct eye contact and soft lip sync. The red border of the drape shifts naturally across her torso while bangles and earrings react with believable secondary motion. At ten seconds create the signature transition entirely through camera movement. The phone rolls clockwise about thirty five degrees while moving slightly closer and lower, turning the composition into a diagonal close up without any digital morph or cut. The background changes naturally in perspective, revealing more of a dark framed window or door structure and a pale ceiling area. From seconds eleven to fifteen maintain this tilted intimate angle as she continues subtle lip sync, tilts her head, briefly touches her hair and gives small emotional expressions directly to camera. Lighting should feel like soft daylight entering from the front and side, gently illuminating the face while the room remains slightly darker. Preserve real exposure variation, fine skin pores, individual hair strands, fabric weave, tiny jewelry reflections, natural motion blur and realistic hand anatomy. Use a smartphone lens look around 26 mm equivalent with moderate depth of field and no artificial cinematic bokeh. No beauty filter, no plastic skin, no face warping, no duplicated jewelry, no extra fingers, no floating fabric, no text, no logos and no watermarks. Music should be an original romantic South Asian pop inspired groove around 90 BPM, with warm hand percussion, soft bass, light melodic strings, subtle electronic rhythm and airy female vocal textures. Match hand gestures, head turns and the camera roll to the beat while keeping the performance spontaneous and natural.

Audio scene

Create a 15 second cinematic martial arts sequence in 16:9 inside an authentic old Asian training hall filled with deep red timber pillars, dark carved walls, geometric lattice windows, a large circular woven ceremonial carpet, low wooden benches, brass vessels, training objects and subtle atmospheric haze. Cool natural daylight enters through the patterned windows while warmer practical interior light softly touches skin and wood. Everything must feel physically photographed on a real set with natural fabric texture, realistic skin pores, imperfect surfaces, believable shadows and subtle airborne dust. Open with a wide eye level view of an athletic woman dressed in a simple black sleeveless training top and dark trousers crouching defensively on the polished wooden floor. Another female fighter briefly crosses close to camera as a long narrow steel weapon sweeps horizontally through the extreme foreground. Use authentic motion blur and shallow depth as the metal passes dangerously close to faces without making contact. Move into a medium view of the black dressed fighter calmly tracking the weapon while the circular carpet and ornate hall remain spatially consistent behind her. Cut to an intense close portrait of a rugged hooded male martial artist seated completely still in pale textured robes. A polished metal blade passes only inches in front of his eyes and cheek. He responds only with tiny eye movement and controlled breathing. Capture realistic eyelashes, beard texture, cloth fibers, reflected window light in his eyes and a softly defocused background. No exaggerated reactions. Return to a wide symmetrical composition revealing the hooded man seated beside the circular carpet as the long weapon cuts across the room toward the female fighter. She drops naturally to the floor, twists underneath the moving weapon and uses one arm for balance. Her body weight, momentum, clothing folds and floor contact must obey real physics. Continue into a low angle shot as she rolls into a deep backward bridge and lets the weapon pass just above her torso. At the action peak perform one fast motivated camera roll toward the timber ceiling as the steel weapon streaks across the image, then immediately return upright. The woman rises fluidly into a strong standing position, lifts one arm above her head and redirects the attack with a precise defensive movement. Finish on a wider centered shot of her holding a grounded martial arts stance against the glowing doorway while the hooded man remains calm in the distant background. Use restrained handheld camera energy mixed with stable cinematic framing, realistic lens breathing, gentle focus transitions, subtle exposure changes and naturally imperfect motion. Avoid impossible anatomy, floating limbs, rubber movement, excessive slow motion, artificial sharpening, plastic skin, fantasy energy effects or supernatural physics. Soundtrack should feature tense cinematic percussion, deep hand drums, subtle low strings and restrained traditional Asian instrumental textures. Synchronize brief metallic swishes, cloth movement, foot contact and sharp air movement with the choreography. Let the music build progressively, peak during the backward dodge and camera roll, then end with one controlled low percussion hit as the final stance lands. No vocals and no dialogue.

Audio scene

Native 16:9 photorealistic premium vertical streaming micro-drama, polished television-drama realism, actor-first framing, natural skin texture, clean readable faces, realistic modern apartment at night. A married couple is inside their living room. The wife stands near the sofa holding her husband's black smartphone in her right hand. The husband enters from the background and stops when he notices the phone in her hand. 0-5s: Begin with a medium close-up of the wife. She casually looks at the phone. A message notification appears from a contact belonging to her sister, whom she believes is dead. Her expression changes subtly: her small smile disappears, fingers stop moving, eyes remain fixed on the screen. 5-10s: Cut to a tight insert of the phone in her hand, then back to her reaction. She reads silently. Her breathing becomes shallow but controlled. No exaggerated crying or gasping. 10-17s: The husband enters behind her. Use a restrained medium two-shot. He sees what she is holding and immediately stops walking. His face becomes tense, but he tries to appear normal. 17-23s: Medium close-up on the wife as she slowly raises her eyes toward the husband, not toward camera. She asks quietly: “Why is my dead sister texting you?” Natural lip sync, low controlled voice. Her eyes begin to water but no tears fall. 23-27s: Reverse medium close-up of the husband looking toward her. He does not answer immediately. His jaw tightens, he swallows once, and his eyes briefly shift toward the phone. 27-30s: Very slow push-in toward the husband as he says quietly: “She’s not dead.” Immediately cut to a tight reaction close-up of the wife. Her eyes widen slightly, completely stunned. End with the wife frozen, phone still in her right hand, staring toward the husband, while the final line hangs in silence. Hold the final reaction for approximately 0.7 seconds for a clean cliffhanger cut. Camera style: eye-level, restrained static coverage with only one slow push-in near the reveal. Maintain consistent screen direction and eyelines. Clean soft motivated apartment lighting, visible eyes, natural contrast, recognizable environment. Audio: quiet apartment room tone, subtle phone notification sound, clear dialogue, almost complete silence after “She’s not dead.” Avoid fashion posing, excessive cinematic haze, dramatic orbiting camera, exaggerated acting, random extras, identity drift, changing wardrobe, phone changing hands, eyeline flips, anatomy errors, unnecessary camera movement, promotional movie-poster look.

Audio scene

POV first-person perspective. The character is an ultra-realistic tiny human, only the size of a human palm. He is **not** a 3D model, cartoon, CGI, toy, figurine, or miniature statue. He looks like a real miniature person with lifelike skin, natural hair, realistic facial features, authentic clothing, and convincing human movements. Everything in the scene is real, including the table, the owner's hands, the ice cream cone, the lighting, and the environment. The tiny man is sitting on a real wooden table. The owner reaches in with a real human hand and gently pokes the tiny man's cheek with a finger. The tiny man immediately looks annoyed, frowns deeply, pushes the giant finger away with both hands, turns his back, puffs out his cheeks, and sulks with crossed arms. The owner then brings a real ice cream cone into the frame and gently taps the top of the tiny man's head with it. The tiny man glances sideways at the ice cream. His eyes instantly light up with excitement. His expression changes from grumpy to delighted. He quickly turns around, eagerly grabs the giant ice cream cone with both hands, happily licks the melting ice cream, leaving a little cream around his lips. The owner gently pats the tiny man's head with a real finger. The tiny man smiles politely with a grateful and satisfied expression, gives a respectful nod, then immediately hugs the giant ice cream cone and runs toward the corner of the real table. He sits down happily and continues eating the ice cream with obvious enjoyment. Hyper-realistic tiny human illusion, miniature real person, macro photography, true-to-life human proportions, natural facial expressions, realistic skin texture with visible pores, lifelike eyes, authentic clothing physics, smooth natural body movements, perfect scale interaction with giant real objects, cinematic realism, physically accurate lighting, natural contact shadows, seamless integration into the real environment, shallow depth of field, handheld macro camera, one continuous shot, no cuts, no dialogue, no subtitles.

這個模型表現如何？

公開資料已經存在，本站仍須完成安全生成支援才能接受輸入。

它擅長什麼
這是關於 Google 的 Gemini 3.8 Flash-Lite TTS 的模型資訊頁。
適合哪些使用者
適合關注可用狀態，或需要現在完成創作任務的使用者。
為什麼要在 Kyeo AI 使用
開放前會核對輸入、輸出用量及計費邊界，避免不確定請求扣取積分。
介面資料已公布。生成入口暫時關閉。同類可用模型列於下方。
Gemini 3.8 Flash-Lite TTS 可用狀態Gemini 3.8 Flash-Lite TTS 替代模型Gemini 3.8 Flash-Lite TTS 生成支援Gemini 3.8 Flash-Lite TTS 輸入要求

使用前須知

目前不接受新生成請求。
輸入及附件控制項未開放。
沒有承諾恢復日期。

同類模型推薦

還沒決定該用哪個嗎？順便比較一下這幾個同類候選模型。

ElevenLabs Dialogue V3

ElevenLabs Dialogue V3 目前提供逐行對白生成：每行填入台詞並選擇聲音，再依台詞總字元數計費。它不是即時語音代理；官方也提醒可能需要產生多個版本才能選到可用結果。本站建議先用樣本確認角色切換、音訊標籤、語言代碼、輸出格式或長腳本完整性。

查看詳情

ElevenLabs Turbo 2.5

ElevenLabs Turbo 2.5 是 Kyeo 目前保留的單人文字轉語音工作流程，可選擇站內聲音，並調整 stability、similarity_boost、style、speed、timestamps、前後文與 language_code。它仍可提交，但 ElevenLabs 目前已建議改用 Flash v2.5；頁面不把低延遲、語言支援、時間戳記精度、聲音授權或輸出品質視為 Kyeo 的保證。

查看詳情

Suno Sounds

Suno Sounds 是 Kyeo AI 的文字生成聲音入口，可選擇 V6 / V6_MINI / V6_WILD、單次音效或循環取向，並可設定 1–300 BPM 與 Key。官網 Beta 功能、樣本數量與訂閱權利不會自動等同於站內結果。

查看詳情

常見問題

資訊來源

Gemini 3.8 Flash-Lite TTS 已公布介面資料，目前在 Kyeo AI 暫不可生成。

模型狀態複核
串接

本次複核確認資料已公布及本站暫停呼叫的狀態。