Mokhanni oa ho nka liphetoho
Ho sebelisoa ka khoebo
380+ li-models
Ha ho letšoao la metsi
Ha ho hlokahale ho ngolisoa
Mofuta:
+ GPT-5, Claude, Gemini
Kopitsa video efe kapa efe le motho ka eona - AI e latela li-keypoints tsa 33 tsa'mele ka lebokoseng le u fa video ea skeleton overlay le JSON ea libaka tse kopaneng bakeng sa lebokose le leng leng. Ha ho na mocap suit, ha ho na li-markers, ha ho na ho hlophisoa. Ho nka li-marker tsa kamera ka MediaPipe.
Tracking 33'mele joints ka ho feletseng ka pheletsong e ngoe le e ngoe...
Likhetho tse tsoetseng pele
Bo_lemo
Tokens e tlase.
Fumana Token e eketsehileng
❤️ U rata Free.ai? Reka le metsoalle ea hau!
Register ho fumana sehokela sa ho u joetsa le ho fumana li-token tse 25 000 ka motsoalle.
U batla ho feta?
Ngola mahala bakeng sa 30K tokens / letsatsi + 10K bonus
Ngola mahala
Ho sebetsana le kopo ea hau...
Ho kenya video, AI e tlosa 3D body pose ka frame ka ho sebelisa MediaPipe. Fumana video ea skeleton overlay hammoho le per-frame keypoints JSON bakeng sa animation, ho hlahloba lipapali, kapa biomechanics. Ha ho mahala, ha ho na li-markers, ha ho na mocap suit.
Mokhoa oa ho sebelisa Mokhanni oa ho nka liphetoho
1
Ke eng eo u e kentseng?
Tlatsa mongolo, kenya faele, kapa hlalosa seo u se batlang. Ha ho hlokahale ak'haonte.
2
Tobetsa ho theha
AI ea rona e sebetsana le kopo ea hau ka metsotsoana ka ho sebelisa li-models tse ntlehali tsa open-source.
3
Tlosa & & arolelana
Kopitsa, kenya kapa arolelana litlamorao tsa hau. Haholo-holo bakeng sa ho sebelisana le batho ba bang le ho rekisa.
Senya sesebelisoa ka API
E-ba le sesebelisoa sena ka ho iketsetsa ho tloha ho kotloloho. OpenAI-compatible REST endpoint, Bearer-token auth, ha ho hlokahale SDK e eketsehileng. Litheko tsa token li lumellana le interface ea webosaeteng.
curl -X POST https://api.free.ai/v1/video/generate/ \
-H "Authorization: Bearer sk-free-..." \
-H "Content-Type: application/json" \
-d '{"prompt": "A cat playing piano", "duration": 4}'Mokhanni oa ho nka liphetoho — FAQ
Ho kenya video le motho ka sehloohong le AI e nka li-joints tsa'mele oa 33 - moelelo, li-shoulders, li-elbows, li-pols, hips, li-knee, li-ankles, hammoho le matsoho le maoto - ka sehloohong se seng le se seng. U tla fumana video ea ho kopanya skeleton le faele ea JSON le li-coordinates tsa ho kopanya ka sehloohong. Ha ho na mocap suit, ha ho na li-markers, ha ho na mohato oa ho kopanya, kamera e le 'ngoe e sebetsa hantle.
Ho na le li-3D animations tsa li-character ka Blender / Unity / Unreal (ho khutlela ho armature e kopantsoeng), ho etsa lipapali / ho bapala / ho etsa li-martial-arts, ho theha li-overlays tsa ho tlosa liphoso, ho ruta li-ML models ka data ea ho tsamaea, kapa ho bona feela li-patterns tsa ho tsamaea ka nako.
MediaPipe Pose Landmarker (Google, Apache 2.0). E fana ka li-keypoints tsa'mele tse 33 ka frame ka li-coordinates tse tloaelehileng tsa 2D + Z e behiloeng (bophahamo bo amanang le kamera) + lipalo tsa ho bona li-keypoint. E sebetsa ka ho feletseng ho CPU ka hona GPU e lula e le mahala bakeng sa li-generations tse ling tsa hau.
Ke 2.5D - 2D ea 'nete + Z e behiloeng ho tsoa ho kamera e le 'ngoe. Ho nka 3D ho hloka li-camera tse ngata tse synchronized bakeng sa triangulation (ho phethela FreeMoCap / OptiTrack / Vicon). Bakeng sa lidansi tsa TikTok, ho hlahloba lipapali, ho hlalosa animation, kapa ho sebetsa ka li-camera tse le' ngoe, tlhahiso ea MediaPipe e ntle. Re tla eketsa sesebelisoa sa li-camera tse ngata ka mor'a moo bakeng sa basebelisi ba hlokang 3D ea 'nete.
200 tokens ka metsotsoana ea video ea ho kenya (e thehiloe ho 500 tokens). Klip ea 10-second e theko e boima ea 2,000 tokens; Klip ea 60-second e theko e boima ea 12,000. Tokens tsa mahala tsa letsatsi le letsatsi li koahela li-clips tse khuts'oane tse'maloa ka letsatsi; basebelisi ba nang le lengolo-tsoibila ba fumana 30K / letsatsi.
Real-time-ish: ka liforeimi tsa 30-50 ka sekoahelo sa rona. Video ea 1-minute 30 fps e sebetsa ka metsotsoana ea 30-60 ho tloha qalong ho isa qetellong ho kenyeletsoa ho kenya + ho etsa. Livideo tse telele li nka nako e telele.
MP4, MOV, WebM, AVI, MKV, le li-formats tse tloaelehileng tsa video — leha e le eng eo ffmpeg e ka e decodeng. Max upload 100 MB. Resolution ha e na bothata bo boholo; moelelo oa pose kahare o theola li-downsamples bakeng sa lebelo.
MediaPipe Pose Landmarker e latela motho e mong ka frame (e moholo ka ho fetisisa). Bakeng sa ho latela batho ba bangata re hloka mofuta o fapaneng (RTMPose, YOLOv8-pose). Haeba u sebelisa batho ba bangata, kenya maikutlo ka /contact/ — re tla leboha haeba u ka e eketsa e le sesebelisoa se seng.
Li-joint tse hlahang li tlalehiloe ho ~5-10 px ka 720p frame; li-joint tse koahetsoeng (leoto ka morao, maoto kantle ho frame) li tlametsoe ka liphetho tse tlase tsa ho bona hore u ka li tlosa. Ho tlosa li-frames ka li-pipeline tsa hau (Kalman / Savitzky-Golay) ho hloekisa ho se nang letho.
Ha se kahare ho sesebelisoa sena ka ho toba kajeno — JSON ke data ea keypoint e sa phekoloeng. U ka fetola ntle le marang-rang ka ho sebelisa libuka tsa libuka joalo ka `aniposelib` kapa `pose-format`. Re nahana ka ho romella "Mocap → BVH/FBX" sesebelisoa sa ho latela — kenya upvote ho /contact/ haeba u e batla.
Ho sebetsana kapele, li-keypoints li tlosoa, ebe video ea ho kenya e tlosoa. Mofuta oa ho tsoa oa skeleton-overlay le JSON li bolokiloe bakeng sa ho fela ha ho arolelana-link (24 h anonimous / 7 d paid). Ha ho sebelisoa ho ithuta. /privacy/ bakeng sa melao e felletseng.
E-na — POST faele ea `video` e nang le likarolo tse ngata ho /v1/video/motion-capture/. E khutlisa {video_url, json_url, duration_s, tokens, share_url}. Bearer auth (sk-free-…) e u fa 10,000 tokens / khoeli e mahala. Curl mohlala ho /api/.
U tla lekola eng ka sesebelisoa sena?