AI Motion Capture
Kushandiswa kwekutengesa OK
380 + mamodheru
Hapana mvura
Hapana kumbobvira kushanyira
Model:
+ GPT-5, Claude, Gemini
Upload chero video ine munhu mu esso — AI tracks 33 chimiro keypoints per frame uye anopa iwe skeleton overlay video pamwe JSON yezvimwe nzvimbo dzese dzese. No mocap suit, hapana mapepa, hapana kuongorora.
Kutevera 33 muviri zvifukidziro pamusoro peese frame…
_Zvirongwa
Chikamu
Tokens iri pasi.
Get More Tokens
❤️ Love Free.ai? Tinya pano kuti utore Free.ai!
Sign up to get a referral link and earn 25,000 tokens per friend.
Uchida zvakawanda?
Sign up for free for 30K tokens/day + 10K bonus
Sign Up Free
Kugadzirisa yako mibvunzo...
Upload a video, AI inotora 3D body pose per frame using MediaPipe. Get back a skeleton overlay video plus a per-frame keypoints JSON for animation, sports analysis, or biomechanics. Free, no markers, no mocap suit.
Maitiro ekuisa AI Motion Capture
1
Sarudza yako input
Tinya meseji, wedzera faira, kana kuti nyora zvaunoda. Hapana account yaunoda.
2
Tinya kuumba
Our AI inoongorora yako mibvunzo mumasekondi nekushandisa yakanakisa open-source mamodheru.
3
Dhawunirodha & shandisa
Dhawunirodha, kopa kana kugovera yako mhinduro. Yemahara yemunhu uye yekutengesa kushandiswa.
Usashandisa iyi chirongwa kuburikidza neAPI
Automatize iyi chirongwa kubva yako pachako code. OpenAI-inowirirana REST endpoint, Bearer-token auth, hapana zvishoma SDK zvinodiwa. Token mutengo kusangana web interface.
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}'AI Motion Capture — FAQ
Kudonha mu video nemunhu mu frame uye AI tracks 33 muviri zvifukidziro — pfungwa, makumbo, makumbo, makumbo, hips, makumbo, makumbo, pamwe nemano uye makumbo — mu chero frame. Unogona kuwanazve skeleton overlay video pamwe neJSON faira ne per-frame joint coordinates. No mocap suit, hapana mapepa, hapana kuongorora step, imwe kamera inoita zvakanaka.
Dzivisa 3D animations yechiratidzo muBlender / Unity / Unreal (re-target kune rigged armature), gadzira mitambo / danho / martial-arts tekinoroji yekuongorora, gadzira overlays yekugadzirisa, dzidzisa ML mamodheru pa data rekufamba, kana kungoona mafambiro ekufamba nenguva.
MediaPipe Pose Landmarker (Google, Apache 2.0). Inoburitsa 33 body keypoints paframe mu normalized 2D coords + a estimated Z (relative depth from the camera) + per-keypoint visibility scores. Inofamba zvachose pa CPU saka GPU inoramba yakazara kune mamwe marudzi ako.
Izvi zvinonzi 2.5D — true 2D + inofungidzirwa yakaenzana Z kubva kune imwe kamera. Real 3D motion capture inodiwa akawanda akaenzana kamera e triangulation (FreeMoCap / OptiTrack / Vicon nzira). Kuti TikTok madansa, mitambo yekuongorora, animation reference, kana chero imwe kamera workflow, MediaPipe's output ndeyekubudirira. Tichawedzera multi-camera tool gare gare kune vashandisi vanoda true 3D.
200 tokens pamusi wekutanga wevhidhiyo (inosvika 500 tokens). 10-sekondi clip inodhura 2,000 tokens; 60-sekondi clip inodhura 12,000. Daily-pool yemahara tokens inosvika pazvimwe zvishoma zvevhidhiyo pazuva; vanopinda vanowana 30K/zuva.
Real-time-ish: pamusoro 30-50 maframes pasecond pabhokisi redu. 1-minute 30 fps video maitiro mu 30-60 masekondi end-to-end kusanganisira kurodha + render.
MP4, MOV, WebM, AVI, MKV, uye akawanda anowanzoitika mavhidhiyo mafomati - chero chinhu ffmpeg inogona decode. Max upload 100 MB. Resolution haina chekuita zvakanyanya; pose model inowedzera simba rekushandisa.
The MediaPipe Pose Landmarker tracks ONE person per frame (the most prominent one). For multi-person tracking we'd need a different model (RTMPose, YOLOv8-pose). If your use case is multi-person, file an idea via /contact/ — happy to add it as a separate tool.
Zviri pachena zvidimbu zvinotevera kusvika ~ 5-10 px pane 720p frame; zvakavharwa zvidimbu (chiuno pasi peshure, chiuno off-frame) zvinodzokororwa neyakaderera kuratidzwa kwezvinyorwa kuti ugone kudzokorora. Kuchenesa pakati pemitsara munzira yako yekukwira (Kalman / Savitzky-Golay) inobvisa zvimwe.
Hapana zvakananga kubva kune iyi tool nhasi — JSON ndiyo raw keypoint data. Unogona kushandura offline using libraries like `aniposelib` or `pose-format`. Tichifunga nezvekutumira "Mocap → BVH/FBX" follow-on tool — tumira upvote pa /contact/ kana uchida.
Kugadziriswazvezve, makeypoints anobuda, wozobvisawo video. Iyo skeleton-overlay output uye JSON inochengetwa kwenguva yekuti link ienderere mberi (24 h anonymous / 7 d paid). Hapana kushandiswa kwedzidzo. /privacy/ for the full policy.
Yeah — POST a multipart `video` file to /v1/video/motion-capture/. Returns {video_url, json_url, duration_s, tokens, share_url}. Bearer auth (sk-free-…) inokupa iwe 10,000 tokens / month free. Curl muenzaniso pa /api/.
Ungaishandisa sei iyi chirongwa?