Reframe any video to vertical, automatically.

One API call turns landscape footage into a 9:16 clip that follows the active speaker. No editor, no manual cropping.

Bash

# Reframe a landscape video to 9:16, following the active speaker
curl -X POST https://naralens.com/api/v2/reframe \
  -H "Authorization: Bearer $NARA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"url": "https://example.com/video.mp4"}'

Get started in three calls

No SDK to install. Just HTTP: submit, poll, download.

Create an API key

Bash

Authorization: Bearer sk-...

Submit a video

POST a public URL. You get back a job_id immediately. The pipeline runs async.

Bash

curl -X POST https://naralens.com/api/v2/reframe \
  -H "Authorization: Bearer $NARA_API_KEY" \
  -d '{"url": "https://example.com/video.mp4"}'

Poll & download

Poll one endpoint until status is completed, then download the 9:16 clip.

Bash

curl https://naralens.com/api/v2/reframe/$JOB_ID \
  -H "Authorization: Bearer $NARA_API_KEY"
# → { "status": "completed", "download_url": "..." }

Read the full reference

Capabilities

One API for video understanding

Built on a serverless GPU pipeline: ingest, detect, track, and reframe, exposed as simple async endpoints.

Speaker-aware reframe

Core

Auto-crop landscape video to 9:16 that follows whoever is talking, with EMA-smoothed panning and scene-cut snapping.

Active speaker detection

Core

Frame-by-frame scores for who is speaking, powered by TalkNet-ASD over tracked faces.

Scene-aware cropping

Core

Scene-cut detection keeps the crop re-centering cleanly across cuts instead of drifting between shots.

Async REST API

Core

Submit a URL, poll a single endpoint, download the result. Plain HTTP, no SDK required.

Pay as you go

Free tier

1 credit = 1 second of video, $1 = 1 minute. Start with 300 free credits, top up with Stripe.

LLM reframe + diarization

Optional

Opt in to hybrid LLM crop planning and speaker diarization for tricky multi-face scenes.