Explain (LLM)

Leverage AI language models to explain, summarize, translate, and simplify text in supported Nigerian languages. Explain is synchronous — the response contains the processed text directly, with no job to poll.

Actions

Action	What it does
`explain`	Clarify the meaning of the text for an interested layperson.
`summarize`	Condense the text to its most important points.
`translate`	Translate into `target_language`, preserving tone and nuance.
`simplify`	Rewrite in simpler, clearer language while preserving the meaning.

Languages & codes

Explain is a text/LLM feature, so — unlike the speech modules — every Nigerian language stands on its own, and Igbo (ig) is live today (even where TTS and ASR still list it as coming soon).

Language	Code
English	`en`
Yoruba	`yo`
Igbo	`ig`
Hausa	`ha`
Nigerian Pidgin	`pcm`

Explain text

POST/dev/v1/explain

Requires the TRANSCRIPTIONS_WRITE scope.

Request parameters

Parameter	Type	Status	Description
`text`	string	Required	The text to explain, summarize, translate, or simplify. 1–12,000 characters.
`language`	string	Required	Language code of the input: `en`, `yo`, `ig`, `ha`, or `pcm`.
`action`	string	Required	`explain`, `summarize`, `translate`, or `simplify`.
`target_language`	string	Optional	Output language for a translation. Required when `action` is `translate`, and must differ from `language`; leave it out for any other action.
`max_words`	integer	Optional	Soft target for response length in words, 1–1000. Default 250.

max_words is a soft target, not a hard cap

The response lands close to the requested length and always finishes its sentences — it is never truncated mid-thought. Billing meters the model's real token usage (tokens_used), so the same max_words costs more credits_used for Nigerian languages (yo, ig, ha) than for English, whose tokenizer is more compact.

cURL

curl -X POST https://api.myvoicemaker.ai/dev/v1/explain \
  -H "Authorization: Bearer vmk_live_a1b2c3d4e5f6_8h9i0jKLmnOPqrSt" \
  -H "Content-Type: application/json" \
  -d '{
    "text": "Èkó ní gbogbo ìgbà tí a bá rí àwọn olùkùnrin méjì...",
    "language": "yo",
    "action": "translate",
    "target_language": "en",
    "max_words": 250
  }'

Python

from myvoicemaker import VoiceMaker

client = VoiceMaker(api_key="vmk_live_a1b2c3d4e5f6_8h9i0jKLmnOPqrSt")

result = client.explain.process(
    text="Èkó ní gbogbo ìgbà...",
    language="yo",
    action="translate",
    target_language="en",
    max_words=250,
)
print(result.result)        # processed text
print(result.tokens_used)   # e.g. 145
print(result.credits_used)  # e.g. 145

JavaScript

import { VoiceMaker } from "myvoicemaker";

const client = new VoiceMaker({ apiKey: "vmk_live_a1b2c3d4e5f6_8h9i0jKLmnOPqrSt" });

const result = await client.explain.process({
  text: "Èkó ní gbogbo ìgbà...",
  language: "yo",
  action: "translate",
  target_language: "en",
  max_words: 250,
});
console.log(result.result);       // processed text
console.log(result.tokens_used);  // e.g. 145
console.log(result.credits_used); // e.g. 145

Response — 200 Success

JSON

{
  "requestId": "8f1c2d3e-4a5b-6c7d-8e9f-0a1b2c3d4e5f",
  "id": "explain_1a2b3c4d5e6f7g8h9i0j",
  "action": "translate",
  "language": "yo",
  "result": "In Lagos, whenever we come across two men...",
  "tokens_used": 145,
  "credits_used": 145,
  "created_at": "2026-03-03T10:30:00Z"
}

action and language echo your request — language is always the input language, even for a translation (the translated text is in result). result is plain prose, with no markdown or preamble.

Billing

Explain bills 1 credit per token, metered on the model's real usage (tokens_used), from your purchased credits only. The estimated cost is reserved up front and settled to the exact usage once the response is ready — you are never charged more than the reserved amount. If your purchased balance is too low, the request fails fast with 402 and insufficient_credits before any work starts.