Minimax Music 2.6 API Documentation

Minimax Music 2.6 API Documentation

Playground

MiniMax Music 2.6 generates complete songs with vocals and instrumentals from text prompts and lyrics. Supports instrumental-only mode, auto lyrics generation, structure tags for song arrangement, and configurable audio quality. Ready-to-use REST inference API, best performance, no coldstarts, affordable pricing.

Features

MiniMax Music 2.6 generates complete songs with vocals and instrumentals from a text prompt and lyrics. Describe the genre, mood, instruments, and tempo — the model composes and sings a full track with your lyrics, with support for structure tags, instrumental-only mode, auto lyrics generation, and configurable audio quality up to 256kbps/44.1kHz.

Why Choose This?

• Complete song generation Produces full vocal and instrumental tracks from a single prompt and lyrics input — not just clips or loops.

• Structure tag support Use tags like [Verse], [Chorus], [Bridge], and more in your lyrics to precisely control song arrangement and flow.

• Auto lyrics generation Leave the lyrics field empty and the model will generate lyrics automatically from your prompt description.

• Instrumental mode Enable is_instrumental to generate a backing track without any vocals — ideal for music beds, soundtracks, and production use.

• Configurable audio quality Adjust bitrate (up to 256kbps) and sample rate (up to 44.1kHz) to match your delivery requirements.

• Wide genre support Pop, rock, folk, electronic, orchestral, lo-fi, and more — all controllable through your prompt.

Parameters

Lyrics Structure Tags

Use these tags in your lyrics to control song arrangement:

[Intro], [Verse], [Pre Chorus], [Chorus], [Post Chorus], [Interlude], [Bridge], [Outro], [Hook], [Build Up], [Break], [Transition], [Inst], [Solo]

How to Use

1. Write your prompt — describe the genre, instrumentation, BPM, mood, and production style. Use the Prompt Enhancer for better results.
2. Enter your lyrics — add structure tags to control arrangement (e.g., [Verse], [Chorus]). Leave empty for auto-generated lyrics.
3. Set bitrate and sample_rate (optional) — adjust for your target delivery format.
4. Enable is_instrumental (optional) — check this to generate a backing track without any vocals.
5. Submit — generate and download your track.

Example Prompt

warm acoustic folk, fingerpicked guitar, soft female vocal, cozy afternoon feeling, gentle reverb, intimate recording, storytelling mood, BPM 72

Example Lyrics

[Verse] Sunlight through the window pane Coffee getting cold again

[Chorus] Time moves slow on days like these Carried softly by the breeze

Pricing

Just $0.15 per song (up to 5 minutes).

Best Use Cases

• Songwriting & demos — Turn lyrics and a style description into a full demo track instantly.
• Content creation — Generate original songs or background music for videos, podcasts, and social media.
• Music production — Use instrumental mode to create backing tracks for further mixing and production.
• Custom soundtracks — Produce original music for games, ads, and film with precise style control.
• Creative exploration — Rapidly prototype different musical styles and arrangements from the same lyrics.

Pro Tips

• Be specific in your prompt — include genre, BPM, key instruments, vocal style, and production cues like “gentle reverb” or “intimate recording.”
• Use structure tags in your lyrics to define verse, chorus, and bridge sections for a more polished arrangement.
• Leave lyrics empty to let the model auto-generate them from your prompt — great for quick ideation.
• Try instrumental mode first to validate the musical feel before committing to a full vocal version.
• Higher bitrate and sample rate settings deliver better audio quality for professional delivery.

Notes

• Both prompt and lyrics are required fields; leave lyrics empty to trigger auto-generation.
• Lyrics length: 10–3000 characters. Prompt length: up to 2000 characters.
• Generated audio URLs expire after 24 hours — download your output promptly.
• Please ensure your content complies with MiniMax’s usage policies.

Related Models

• MiniMax Music 2.5 — Previous generation MiniMax music generation model.
• MiniMax Speech 2.8 HD — High-definition text-to-speech generation from MiniMax.

Authentication

For authentication details, please refer to the Authentication Guide.

API Endpoints

Submit Task & Query Result

set -euo pipefail

export WAVESPEED_API_KEY="your-api-key"

REQUEST_BODY=$(cat <<'JSON'
{
  "prompt": "A cinematic ocean wave at sunrise, highly detailed",
  "bitrate": 256000,
  "sample_rate": 44100,
  "is_instrumental": false
}
JSON
)

# 1. Submit the prediction once. Never retry this request automatically:
# a retry after an unclear failure can create a second, billed task.
CHECK='The task may already exist. Check History or the predictions list before submitting again.'
if ! SUBMIT_RESPONSE=$(curl --silent --show-error --fail-with-body \
  --connect-timeout 10 --max-time 60 \
  -X POST "https://api.wavespeed.ai/api/v3/minimax/music-2.6" \
  -H "Authorization: Bearer ${WAVESPEED_API_KEY}" \
  -H "Content-Type: application/json" \
  -d "${REQUEST_BODY}"); then
  printf 'Submit failed. %s\n' "${CHECK}" >&2
  exit 1
fi

TASK=$(printf '%s' "${SUBMIT_RESPONSE}" | jq 'if type == "object" and has("data") then .data else . end')
PREDICTION_ID=$(printf '%s' "${TASK}" | jq -r '.id // empty')
if [ -z "${PREDICTION_ID}" ]; then
  printf 'Submission response did not contain a prediction id. %s\n' "${CHECK}" >&2
  exit 1
fi
RESULT_URL="https://api.wavespeed.ai/api/v3/predictions/${PREDICTION_ID}/result"

# 2. Poll until the prediction reaches a final status.
BODY_FILE=$(mktemp)
trap 'rm -f "${BODY_FILE}"' EXIT
while true; do
  # Transient errors (429, 5xx) are retried, honoring Retry-After.
  HTTP_CODE=$(curl --silent --show-error --fail-with-body \
    --connect-timeout 10 --max-time 60 --retry 5 --retry-delay 2 \
    --output "${BODY_FILE}" --write-out '%{http_code}' \
    "${RESULT_URL}" \
    -H "Authorization: Bearer ${WAVESPEED_API_KEY}" || true)
  case "${HTTP_CODE}" in
    2??) ;;
    000|429|5??)  # dropped connection or temporary error: keep polling
      printf 'Task %s: HTTP %s, polling again in 10s\n' "${PREDICTION_ID}" "${HTTP_CODE}" >&2
      sleep 10; continue ;;
    *) printf 'Task %s: HTTP %s\n' "${PREDICTION_ID}" "${HTTP_CODE}" >&2
      cat "${BODY_FILE}" >&2; exit 1 ;;
  esac
  RESULT=$(jq 'if type == "object" and has("data") then .data else . end' "${BODY_FILE}")
  STATUS=$(printf '%s' "${RESULT}" | jq -r '.status // empty')

  case "${STATUS}" in
    completed) printf '%s\n' "${RESULT}" | jq '.outputs'; break ;;
    failed|cancelled|timeout|deleted)
      printf 'Task %s ended with status %s\n' "${PREDICTION_ID}" "${STATUS}" >&2
      printf '%s\n' "${RESULT}" | jq . >&2; exit 1 ;;
    *) sleep 2 ;;
  esac
done

Parameters

Task Submission Parameters

Request Parameters

Every model also accepts the global request options enable_sync_mode and enable_base64_output; see their pages for details.

Response Parameters

Result Request Parameters

Result Response Parameters