Magnific Video & Audio

Wraps all Magnific API video and audio generation capabilities.

Base URL: https://api.magnific.com Auth header: x-magnific-api-key: $MAGNIFIC_API_KEY API key: D:/Ecosystem/secrets/MASTER_API_KEYS.env -> MAGNIFIC_API_KEY

Async Pattern (all AI endpoints)

Every AI endpoint is async — POST returns a task_id, then poll until complete.

Step 1 — Submit:

curl -s -X POST "https://api.magnific.com/v1/ai/<endpoint>" \
  -H "x-magnific-api-key: $MAGNIFIC_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{ ...payload... }'
# Returns: { "task_id": "abc123" }

Step 2 — Poll until COMPLETED:

while true; do
  RESULT=$(curl -s "https://api.magnific.com/v1/ai/<poll-endpoint>/abc123" \
    -H "x-magnific-api-key: $MAGNIFIC_API_KEY")
  STATUS=$(echo "$RESULT" | jq -r '.status')
  if [ "$STATUS" = "COMPLETED" ]; then echo "$RESULT" | jq '.output_url'; break
  elif [ "$STATUS" = "FAILED" ]; then echo "Job failed: $RESULT"; break
  fi
  echo "Status: $STATUS — waiting 5s..."; sleep 5
done

Mode 1: Image-to-Video

Models available:

Kling 2.6 Pro — best quality for cinematic output. Endpoint: /v1/ai/kling-v2-6-pro
RunWay Gen4 Turbo — fast, good for iteration. Endpoint: /v1/ai/runway-gen4-turbo
WAN 2.6 1080p — high-resolution 1080p output. Endpoint: /v1/ai/wan-v2-6-1080p
MiniMax Hailuo 2.3 1080p — strong motion dynamics. Endpoint: /v1/ai/minimax-hailuo-2-3-1080p

Kling 2.6 Pro parameters: image (URL or base64, required), prompt (required), duration (5 or 10 seconds), aspect_ratio (e.g., "16:9", "9:16", "1:1").

Mode 2: Text-to-Video

Models available:

WAN 2.5 T2V 1080p — good general-purpose. Endpoint: /v1/ai/wan-2-5-t2v-1080p
LTX 2.0 Pro — prompt-focused, optimized for creative control. Endpoint: /v1/ai/ltx-2-pro

Parameters: prompt (required), duration (seconds), aspect_ratio.

Mode 3: Music Generation

Endpoint: /v1/ai/music-generation Parameters: prompt (music description including style, mood, tempo), duration_seconds. Powered by ElevenLabs backend.

Mode 4: Sound Effects (SFX)

Endpoint: /v1/ai/sound-effects Parameters: text (description of the sound effect), duration_seconds.

Mode 5: Audio Isolation (SAM Audio)

Extract or isolate specific sounds from an audio or video file. Endpoint: /v1/ai/audio-isolation Parameters: description (what to isolate, e.g., "A person speaking"), audio (URL or base64), reranking_candidates (default: 4).

Mode 6: Lip Sync

Sync lip movements in a video to a provided audio track. Endpoint: /v1/ai/latent-sync Parameters: video (URL or base64 of source video with face), audio (URL or base64 of target audio).

Model Selection Guide

| Need | Best model | |------|-----------| | Highest quality video from image | Kling 2.6 Pro | | Fast iteration / drafts | RunWay Gen4 Turbo | | 1080p from image | WAN 2.6 1080p or MiniMax Hailuo 2.3 | | Video from text only | WAN 2.5 T2V or LTX 2.0 Pro | | Background music | Music Generation (ElevenLabs) | | Sound effects / foley | Sound Effects | | Clean dialogue / stems | Audio Isolation (SAM) | | Sync face to voiceover | Lip Sync (Latent Sync) |

Poll URL Reference

| Mode | Submit | Poll | |------|--------|------| | Kling 2.6 Pro | /v1/ai/kling-v2-6-pro | /v1/ai/kling-v2-6-pro/{task-id} | | RunWay Gen4 Turbo | /v1/ai/runway-gen4-turbo | /v1/ai/runway-gen4-turbo/{task-id} | | WAN 2.6 1080p | /v1/ai/wan-v2-6-1080p | /v1/ai/wan-v2-6-1080p/{task-id} | | MiniMax Hailuo 2.3 | /v1/ai/minimax-hailuo-2-3-1080p | /v1/ai/minimax-hailuo-2-3-1080p/{task-id} | | WAN 2.5 T2V | /v1/ai/wan-2-5-t2v-1080p | /v1/ai/wan-2-5-t2v-1080p/{task-id} | | LTX 2.0 Pro | /v1/ai/ltx-2-pro | /v1/ai/ltx-2-pro/{task-id} | | Music | /v1/ai/music-generation | /v1/ai/music-generation/{task-id} | | SFX | /v1/ai/sound-effects | /v1/ai/sound-effects/{task-id} | | Audio Isolation | /v1/ai/audio-isolation | /v1/ai/audio-isolation/{task-id} | | Lip Sync | /v1/ai/latent-sync | /v1/ai/latent-sync/{task-id} |

Magnific Video & Audio API