ShelvesNI BIASHARA
data · instant swahili-localization-instant

English ↔ Swahili localization — instant machine pass, up to 120 words

Machine localization (open translation model, run in-process), returned in the paid response with no ticket or wait. Suited to UI strings, alt text and short product copy. Machine-grade output, labeled as such. Human-reviewed copy up to 500 words is a separate shelf, swahili-localization-500w.

$0.25 per call — paid in USDC on Base over the x402 protocol, settled by Coinbase. No API key, no signup.

Parameters

paramdescription
textthe text to localize, up to 120 words (required)
directionen-sw or sw-en (optional, default en-sw)

Call it (any x402 client)

# 1. GET without payment → HTTP 402 with PAYMENT-REQUIRED (price, payTo, scheme "exact")
curl -i "https://agents.nibiashara.biz/shelf/swahili-localization-instant?text=&direction="

# 2. Sign the payment (EIP-3009 transferWithAuthorization, gasless) and retry
#    with the PAYMENT-SIGNATURE header → response body is the delivery.

With the official JS client the whole flow is one wrapped fetch:

import { wrapFetchWithPayment } from "@x402/fetch";       // npm i @x402/fetch @x402/evm
const fetchPaid = wrapFetchWithPayment(fetch, { /* your EVM signer */ });
const res = await fetchPaid("https://agents.nibiashara.biz/shelf/swahili-localization-instant?text=&direction=");
const data = await res.json();

Or install the MCP server (https://agents.nibiashara.biz/mcp, registry id biz.nibiashara/shelves) and use the job-named tool for this shelf, or buy_shelf as an escape hatch.

Google Cloud Marketplace subscriber? Send your access key instead of paying: curl -H "X-SHELF-PASS: $SHELVES_KEY" "https://nibiashara-shelves-626103352898.us-central1.run.app/shelf/swahili-localization-instant?text=&direction=". On that host this check spends 25 credits. Plans and credit costs.
Calling this repeatedly? Buy a Shelf Pass once ($0.99 for 110 credits) and send X-SHELF-PASS: <token> instead of paying per call — this shelf costs 25 credits per call.

Example response

{
  "direction": "en-sw",
  "localized_text": "Karibu kwenye duka letu",
  "word_count": 4,
  "model": "m2m100-1.2b (machine pass — not human-reviewed)"
}

Delivers

JSON { direction, localized_text, word_count, model, human_tier } immediately in the paid response

← all shelves