Xpin AI · platform

Everything we’ve built, in one place.

Record it, upload it, annotate it, synthesise it, ship it behind an API. Every product on the voice platform — with the engines and storage underneath — opens from here.

Studio

Products people sign in to and use directly.

Voice Studio

Generate speech and transcribe audio

The Voice Stack workspace. Turn text into speech in any published voice, transcribe recordings into English or Arabic, browse and clone voices, and manage the API keys and usage for your workspace.

Text to speechSpeech to textVoice library
Access
Any workspace account
Opens
voice.xpin.ae/app/home
Open Voice Studio

Emarati Corpus

Bulk-upload and review paired speech data

Builds Emarati Arabic and English training corpora out of audio you already have. Contributors upload paired WAV and transcript files in bulk, reviewers listen and approve or flag each pair, and an approved manifest exports as CSV.

Bulk uploadReview workflowManifest export
Access
Provisioned @xpin.ae accounts only
Opens
av.xpin.ae
Open Emarati Corpus

VeriSpeak

Record a balanced corpus, prompt by prompt

Collects new speech in the browser instead of ingesting existing files. Contributors read one prompt at a time while each take is checked live for clipping, low level and short duration; prompts go out least-recorded-first so coverage stays even. Exports a JSONL manifest with 16 kHz mono WAV.

Guided recordingQuality gatesJSONL export
Access
Contributor and dataset-owner accounts
Opens
corpus.xpin.ae
Open VeriSpeak

ZoéSpeak

Turn ASR output into verified ground truth

Where machine transcripts become training labels. Annotation teams play each clip and correct it word by word against a typed error taxonomy — substitution, mispronunciation, numeric, insertion, deletion, unclear audio — with live word-error rate and a QA gate on every batch.

Word-level editingWER metricsQA gate
Access
Annotation team accounts
Opens
speech.xpin.ae
Open ZoéSpeak

Platform

Shared services that every product builds on.

Middleware API Gateway

Publish, meter and sell any API

One managed front door for every API we expose. Paste a curl command to publish an endpoint, hand out scoped access tokens, enforce rate limits and monthly quotas, and track plans, credits and invoices per customer.

Access tokensQuotasBilling
Access
Super admin and tenant accounts
Opens
voice-middleware.voice-middleware.svc.cluster.local:8400
Open Middleware API Gateway

API Explorer

Try any published API from the studio

A live catalogue of every API published to your workspace, with a request runner and ready-made code snippets in curl, Python and Node. Services registered in the gateway show up here on their own — nothing to wire by hand.

Live runnerCode snippetsAuto-discovery
Access
Any workspace account
Opens
voice.xpin.ae/app/api-explorer
Open API Explorer

API Documentation

Endpoints, parameters and worked examples

The written reference for the public Voice Stack API — synthesis, transcription, voice management and audio streaming — with copyable request and response examples for each endpoint.

ReferenceExamplesArchitecture
Access
Public
Opens
voice.xpin.ae/docs
Open API Documentation

Platform Status

Live health of every component

Probes the API, both speech engines, the gateway, the studio, object storage and the database every few seconds, and reports latency history, per-machine detail and regional availability. Monitors can poll the same data as JSON.

Live probesLatency historyJSON feed
Access
Public
Opens
voice.xpin.ae/status
Open Platform Status

Infrastructure

Engines and storage behind the products.
Adding a product

Register a new tool

This page is generated from a single registry. Add one entry to src/lib/tools.ts with the product’s name, description, category, icon and URL, and it appears here in the right group with its own card. Set an envKey to let each environment point the same tool at a different host without touching code.