Building the future of intelligent work
Our mission is simple to say and hard to do: make frontier AI as dependable as electricity. One platform for chat, image, video, voice and agents — fast enough to trust, safe enough to ship, and priced so any team can start today.
From garage prototype to global platform
Five years, five turning points. The milestones that shaped how we build.
- 2021
Founded in a San Francisco garage
Three researchers leave their labs with one conviction: frontier AI should feel like a tool, not a thesis. The first prototype answers its first prompt in November.
- 2022
AI Chat launches to the public
Our first product ships with multi-model routing built in from day one. 10,000 teams sign up in the first ninety days — the servers barely survive.
- 2023
One platform, every modality
Image, video and voice generation join the platform. The workflow builder arrives, and Tapotik becomes the first suite where modalities share context and credits.
- 2024
Agents & enterprise scale
Autonomous agents ship with human approval gates. SOC 2 Type II, region pinning and zero-retention inference unlock our first Fortune 500 deployments.
- Today
12,000+ teams and counting
48 million AI requests a day across 32 regions, a team of 140 across 14 countries — and the roadmap has never been longer. We're just getting started.
What we optimize for
Four principles that survive every reorg, roadmap and model release.
Ship with care
Speed matters, but craft compounds. We sweat latency budgets, empty states and error copy the same way we sweat model quality.
Default to transparency
Public changelogs, honest pricing, visible status pages. Internally, every metric and every roadmap doc is open to the whole team.
Humans hold the keys
Autonomy with approval gates, AI with audit trails. We build systems where people stay in control of what matters.
Compound learning
Models improve weekly and so do we. Post-mortems without blame, experiments without ego, and a bias for writing things down.
The people behind the platform
A team of researchers, builders and operators from the labs and companies that defined modern AI.
Amara Osei
Co-founder & CEO
Former inference-infra lead; obsessed with making frontier AI boringly reliable.
Jonas Weber
Co-founder & CTO
Built three ML platforms before this one. Believes latency is a feature.
Priya Raghavan
Chief Scientist
Leads model research and evals. Publishes more than she sleeps.
Diego Martins
VP of Design
Designs the interfaces that make streaming tokens feel calm.
Hana Sato
VP of Engineering
Runs the edge network spanning 32 regions — and the on-call rota culture.
Liam O'Connor
Chief Revenue Officer
Took us from zero to 12,000 teams without a single cold deck.
- Active teams
- 0+
- AI requests / day
- 0M
- Uptime SLA
- 0.00%
- Avg. first token
- 0ms
Three hubs, fourteen countries
Remote-first by design. Our hubs anchor time zones — not attendance.
San Francisco
Headquarters
Mission District, CA
London
EMEA Hub
Shoreditch, UK
Singapore
APAC Hub
Tanjong Pagar, SG
Plus teammates working remotely from Berlin, Toronto, Bengaluru, São Paulo, Sydney and nine more time zones.
Ready to build the future of work?
Join 12,000+ teams shipping AI products with Tapotik. Start free — your first thousand credits are on us.
No credit card required · Cancel anytime · SOC 2 Type II