Weekly newsletter audio edition: a year of TTS at 6,000 characters

A weekly newsletter audio edition of 6,000 characters runs to 312,000 characters over 52 issues. A year of narration costs $4.68 on MAI-Voice-2.1-Flash, $6.86 on MAI-Voice-2.1 and $14.82 on Sume TTS at list. A single edition is 9 cents on Flash, 13 cents on MAI-Voice-2.1 and 28.5 cents on Sume.

The bill

For a publisher the annual figure is small enough that the choice is about workflow and listener experience, not budget.

One job per edition

Each edition is one Sume job: 6,000 characters sits well inside the 20,000-character request cap. The result returns a hosted audio URL, which you place in your newsletter's audio block, and the default is MP3 at 44.1 kHz and 128 kbps, a normal podcast-style file.

The Sume cap matters when an issue grows. A 25,000-character special needs two jobs, and the 1,200-second audio limit is the second wall. Split on a section break, not mid-sentence, then join the two files with Timeline audio for a flat $0.01 so readers get one track.

Write the text for the ear before you submit it. Characters count including spaces and punctuation, so tracking links, footers and unsubscribe text that you strip from the spoken version also come off the bill.

Languages and the transcript

Microsoft describes MAI-Voice-2.1 as supporting 23 languages and 26 locales with one voice across all of them. For a multilingual list that is attractive, but it is not available as a Sume voice choice. On Sume, set language per request and pick a voice for that language, and expect a tts_voice_language_warning if they disagree.

If you also want a transcript for the audio page, Sume STT costs $0.01 a minute at list, so a 30-minute edition would be 30 cents. You already have the text, so an STT pass is only worth it for checking the take.

What the table leaves out

Microsoft's figures are the launch prices from 2026-10-01 (Microsoft AI, read 2026-10-05): $22 per 1M characters for MAI-Voice-2.1 and $15 per 1M characters for MAI-Voice-2.1-Flash, which makes up to 45 seconds of audio per call. Sume's figure is its list rate per transcript character, and on Sume spaces and punctuation count toward usage, so count characters of the final text.

The cheaper per-character price does not make the Microsoft route a drop-in swap. A Sume TTS request is an async job whose result carries a hosted audio_url, duration, optional word timings and sentence slices, and that hosted file is what Timeline audio and renders accept. Count the extra work of storing and joining files outside Sume before you call the gap a saving.

Check language coverage first. Microsoft lists 23 languages and 26 locales for MAI-Voice-2.1, and Sume takes a language field that you must set for every non-English transcript. A cheaper model that does not speak your target language is not an option, so settle that before you compare totals.

Make it repeatable

Keep a voice and a speed fixed across the year so the show sounds the same each week. Store the TTS settings next to the newsletter template, and run a dry-run quote on the first edition to see the real character count of what you send.

Label the audio as synthetic voice where your platform requires it. Apple Podcasts and others ask for disclosure of AI voices, which is a publishing matter and not a billing one.

Before you commit to a year, pull one back issue and run it through a dry-run quote. The character count is the only input to the bill, so a quote on a real issue gives you the true annual figure within a few cents. Multiply by 52, add a margin for the occasional long special, and you have a budget line that nobody will ever need to revisit.