Text-to-speech cost for a long podcast script: 5,000 words

5,000 words of narration for a long podcast script cost about $1.4250 on Sume TTS. Sume prices text-to-speech at $0.0475 per 1,000 characters, spaces and punctuation included, and one request accepts at most 20,000 characters, so this script takes 2 jobs.

The character count is an estimate: it assumes about 6 characters per English word including the space. Count your own script with len(text) for the real number.

How is the price built?

The rate comes from Sume's catalog entry sume/tts-1.0: $0.0475 per 1,000 characters, estimated at 1,000 characters, minimum 1, maximum 20,000 per request.

How do I split a script that is over 20,000 characters?

Split on paragraph boundaries into 2 chunks of under 20,000 characters, submit one job per chunk, and join the audio in a Timeline.

import re

def chunks(text: str, limit: int = 20000) -> list[str]:
    out, cur = [], ""
    for para in re.split(r"\n\s*\n", text):
        if len(cur) + len(para) + 2 > limit and cur:
            out.append(cur)
            cur = ""
        cur += para + "\n\n"
    if cur.strip():
        out.append(cur)
    return out

if __name__ == "__main__":
    sample = "First paragraph.\n\nSecond paragraph."
    print([len(c) for c in chunks(sample)])