Ostronaut is out of stealth.

Why this, why now.

Conversation intelligence works for digital teams. Gong became a $7B company on Zoom calls. The same product question — what is the customer actually saying? — has never had an answer on the physical floor, where the conversations that drive most of India’s retail revenue actually happen.

Two things changed. Indic AI finally got good enough to transcribe the code-mixed Marathi · Hindi · English of an Indian retail counter without flattening it to Latin script. And the cost of always-on capture — hardware, networking, storage — dropped to where putting a device on every counter is a rounding error against the revenue it surfaces.

We’ve spent 18 months turning those two changes into a product that works at the noise floor of a real Indian pharmacy at 3pm. Today, we ship it.

What we built.

Six stages, one loop. Each conversation flows from raw audio at the counter to a coachable moment in someone’s queue — without anybody typing anything.

• Capture — Rover, our purpose-built audio hardware in three form factors (Dock for counters, Clip for staff, Vision for camera nodes), tuned for the noise floor of an Indian retail floor.
• Transcribe — our own AI transcription pipeline that handles code-mixed Hindi, Marathi, and English in a single pass, with Devanagari preserved exactly as spoken.
• Diarise — our diarisation AI attributes every turn to the right person, recognises staff by voice embedding across shifts, and keeps customers anonymous.
• Extract — the AI doesn’t hand operators a topic cloud. It surfaces specific business events — stockouts, bounces, missed cross-sells — against a hand-built retail taxonomy.
• Coach — every event becomes a 90-second practice rep, built from the actual conversation, sent to the staff member who lived it before their next shift.

Each stage shipped its own depth before we could call the loop closed. Read the full pipeline →

Where it’s running.

Live across teams building some of the most ambitious operations in the region: pharmacy retail, healthcare sales, D2C consumer brands, B2B sales, and developer tooling. See the customers →

Who backed us.

Y Combinator (W21, via Pragmatic Leaders), Goodwater Capital, and Harvard Management Company. The thesis they backed: the next $7B category-defining conversation intelligence company will be built on the physical retail floor, not on top of more Zoom calls. And it will be built in India, on Indic AI, for the markets where Gong never reached.

The team behind Ostronaut also runs Pragmatic Leaders — the company that has trained more than 10,000 professionals on the practice loop. Six years of evidence on what actually moves capability in front of a customer sits underneath our coaching layer. Read the lineage →

What’s next.

• More verticals: healthcare and quick-commerce conversations are next.
• Next-generation Rover hardware: lower cost, wider acoustic range, longer battery.
• Deeper integrations into the systems your team already uses — CRM, helpdesk, LMS, BI.
• An open MCP server: your AI assistant can already query our intelligence layer directly. See the MCP surface →