Loudly music API tempo, genre and mood vs a Sume Music prompt

Loudly's developer page says its API lets you generate adaptive, royalty-free soundtracks with control over genre, tempo and mood or emotion. Sume's Music Router does it differently: one prompt field of up to 5,000 characters, an optional image_url, and no tempo, genre, mood or duration parameter. You put a number such as "72 BPM" in the prompt and read the result, because the Sume docs call prompt directions "creative directions, not guaranteed output settings".

The Loudly facts below come from its AI Music APIs & Developer Tools for Apps and Games page, read on 2026-10-02. Sume's come from the Music Router and Music 1.0 docs.

What does Loudly's page say the API does?

It lists real-time generation with customizable parameters, a REST API for mobile, web and game engines, text-to-music, stem splitting and remix, and automatic royalty-free licensing for commercial use. The parameters it names are mood or emotion, tempo and genre, plus triggers from system events or user input. Output is finished audio files with full commercial rights, ownership verification and metadata documentation. Pricing is described only as pay-as-you-go, monthly subscription or enterprise.

The page prints no tempo range, no genre list and no price figure, so a comparison of numbers is not possible from it. What can be compared is the shape of the input.

How does a Sume request express the same intent?

Music Router takes POST /v1/music-router/generate with an optional model (omitted means sume/music-auto, which is Lyria 3.5 today; lyria-3.5 and lyria-3-pro are also routable). The Music docs suggest a brief with seven axes: emotion, genre or lineage, tempo as a number, key and mode, two to four instruments with texture, an arc with one named moment, and era or production. These are written into the prompt, not sent as fields.

A request that asks for the tempo, genre and mood Loudly would take as parameters looks like this:

curl -X POST https://api.sume.com/v1/music-router/generate \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: music-bpm-001" \
  -d '{
    "model": "sume/music-auto",
    "prompt": "Hopeful synthwave, 108 BPM, A minor. Analog bass, gated snare, bright arpeggio lead. A 30-second track. Instrumental, no vocals."
  }'

How do the two inputs line up?

How do you get a repeatable tempo out of a text prompt?

Because tempo is prose, the habit that works is to state it once, in the same position, and to pair it with a key and a length so the model has less to guess. The Music docs recommend adding BPM, key and section markers directly to the prompt and treat them as creative direction rather than guaranteed settings. That means a result at 108 BPM is a target, not a contract.

If a downstream step depends on exact tempo, for example cutting video to a beat, listen to the artifact and measure it before you lock the edit. When the length is wrong, trim or loop it with the timeline soundtrack bed instead of regenerating, since a retry is another paid generation.

Loudly's page does not say whether its tempo parameter is a guaranteed value or a hint, so do not assume it is stricter than a prompt without testing.

Which should you pick, and what does Sume not do?

Pick Loudly's style of API if your product needs a tempo or mood to be a typed, validated field, for example a game that maps a slider to a parameter. Pick Sume if the music is one step in a larger media flow and you want the same job, polling and webhook model as the rest of Sume. A Music 1.0 or Router generation is a fixed price per accepted generation, not per second.

• No tempo, genre, mood or duration field; steering is text, and the result must be checked by listening.
• No negative prompt: a non-empty negative_prompt returns HTTP 400 with public_reason=negative_prompt_unsupported. Write exclusions in the positive prompt.
• No seed, so the same prompt can return different music.
• No stem splitting or remix on the Music docs pages.