Maridian turns a voice note into an invoice and a payment link, instantly. No app to download. No behavior to change. Save the contact. Forward the voice note. Done.
Maridian lives inside the messaging thread your business already uses. There is nothing to install, no new interface to learn. The entire workflow happens in a conversation.
Your buyer sends a voice message ordering goods or requesting a service — in any language, any dialect, any accent.
Instead of typing it into a spreadsheet, forward the voice note to your Maridian contact. One tap.
Maridian replies with a structured order summary. If any money field is unclear, it asks you to confirm — it never guesses.
A professional PDF invoice and a request-to-pay link are generated. Forward them back to your buyer. Payment settles via M-Pesa, PIX, or card.
A real voice note. A real invoice. A real payment link. In under ten seconds.
The demo runs on a live WhatsApp thread processing Swahili/Sheng trade speech with background market noise — the same conditions our users operate in every day.
Interactive demo — coming soon
Voice note → Invoice → Payment link
The Smart Contact is the workflow. The Intelligence API is the brain. Together, they convert the dark data of commercial speech into structured financial assets.
A messaging-native business contact that converts speech into structured business actions — invoice, payment link, and records — without changing user behavior.
A specialized engine that enterprises license to understand real-world commercial speech — dialects, slang, code-switching — at scale.
Standard speech models are trained on clean audio — audiobooks, news broadcasts, studio recordings. They break on the messy, multilingual, code-switched reality of how business actually sounds in high-growth markets.
Real commercial speech mixes languages mid-sentence — Swahili and English in Nairobi, Portuguese and regional slang in São Paulo, Pidgin in Lagos. Our engine is built for this reality, not the monolingual ideal.
Trade doesn't happen in quiet rooms. It happens in markets, on construction sites, next to running engines. Our models are trained on the acoustic conditions where real orders are placed — not sanitized studio audio.
We are building a commercial speech dataset that does not exist in the public domain. Every confirmed transaction sharpens the model. This data cannot be scraped from the internet — it can only be earned through deployment.
Every AI system gets things wrong sometimes. Maridian is designed so that when uncertainty arises on a money field, it stops and asks — never guesses.
This is not a feature. It is the primary safety protocol.
If AI confidence on any price, quantity, or amount falls below 97%, the system triggers a verification prompt. The user confirms or re-records. No exceptions.
Starting in the Swahili Belt — where voice notes are the operating layer of wholesale trade — and expanding to Brazil, where "Zap Zap" audio commerce meets instant PIX payments.
Swahili/Sheng trade speech. M-Pesa rails. eTIMS compliance pressure creating digital invoice demand.
Primary Swahili hub. Kariakoo wholesale markets. Highest voice-note ordering density in the region.
"Zap Zap" economy. 4x more voice messages than any other country. PIX instant payments with 56% B2B growth.
Pidgin trade speech. Mexico remittance corridor. The engine is language-agnostic — new dialects are datasets, not rebuilds.
Banks, telcos, and logistics companies spend millions because their software cannot understand customers speaking at full speed in local dialects. The Multilingual Intelligence API changes that.
Automate customer service and fraud detection for customers who speak in vernacular. Reduce cost-to-serve on voice channels by understanding what the customer is actually saying.
Digitize voice-based dispatch centers where drivers communicate through voice notes. Convert chaotic audio into structured route and delivery instructions.
Index and search vast archives of audio content — call center recordings, field operations audio, and broadcast media — making unstructured voice data queryable at scale.
We are onboarding anchor wholesalers and distribution operators for early access in Nairobi, Dar es Salaam, and São Paulo.