We're betting on the human side of the agentic web.
Dart Browser is the only browser that treats your cryptographic identity and your AI agent as things that belong to you — not to a platform. Here's what we believe, and what we're trying to build.
The bet
Every browser made in the last decade has been optimised to serve the platform behind it, not the person in front of it. Chrome serves Google's ad business. Safari serves Apple's ecosystem lock-in. Edge serves Microsoft 365. Even the privacy-focused alternatives are ultimately funded by search engine referrals.
AI makes this worse by an order of magnitude. When the browser becomes an agent — capable of reading, acting, and deciding on your behalf — the question of whose interests it's optimised for stops being philosophical and starts being existential. An agent that sends your browsing context to a company's servers to power their AI is not your agent. It's theirs.
We think the next decade of browser development will be defined by this question: who does the browser serve? We started Dart to build the answer that serves the user.
What we believe
A handful of principles that govern every product decision we make.
Local-first AI is the only honest AI.
If your agent sends your prompts to a third-party server, you have surrendered context about everything you do online to that company. Local inference is not a performance feature — it's a privacy guarantee. We build local-first by default and treat cloud inference as the opt-in exception.
Bring your own model, always.
We charge for the browser. You own the intelligence. Dart is not in the business of selling tokens or marking up inference. Connecting your own API key or a local Ollama instance is a first-class experience, not a workaround.
Hardware cryptography is not optional.
Passwords are a thirty-year-old mistake that phishing has turned into a liability. Hardware-attested identity — where your private key lives in a TPM or Secure Enclave and never leaves — is the floor, not a premium feature. Post-quantum algorithms are on by default because harvest-now-decrypt-later is a real threat model today.
Chromium is the right foundation.
Web compatibility matters. We build on Chromium because it has the widest site support, the most hardened V8 security track record, and the largest external research community. We strip the surveillance infrastructure and add our own layers on top. Purity theatre that breaks half the web serves no one.
Transparency over trust-me.
Every agent action is logged to a local, human-readable audit trail. The telemetry model is off by default. The network requests the browser makes are documented. We do not ask you to trust us — we build tools so you can verify.
Origin story
Dart started from a question that kept coming up in our conversations: "If the AI assistant in your browser were reading everything you do, would you know? And would you care?"
Most people we asked didn't know. The honest ones admitted they'd clicked through terms without reading them because there was no alternative. The arrival of agentic AI in browsers made this question urgent — an agent that can act on your behalf is also an agent that can surveil on someone else's behalf. The discourse shifted in 2025 from "is this real?" to "what do we do about it?", and that's the gap we exist to fill.
We are a small team. We are not pretending we have built a perfect browser. We are claiming that the direction matters — that a browser designed from first principles around the user's sovereignty, rather than retrofitted with privacy features to reduce churn, produces a qualitatively different product. If we're wrong, we'll have built a very good browser. If we're right, the rest of the industry follows.
A small team. Deep expertise. All in.
Dart Browser is being built by a small founding team with deep experience in browser security, cryptographic systems, and applied AI. We're hiring carefully; open roles are on the careers page.
We're a tight founding group with complementary backgrounds across browser engineering, cryptography, applied ML, and product security. Several of us have worked at or founded companies in the browser and security tooling space. We share one conviction: the next decade's browser should be measured by what it does for users, not by how effectively it monetises them.
We're based in Gurgaon, San Francisco, and remote, with a preference for people in Pacific or European time zones so we can overlap for a few hours each day.
Three claims we keep coming back to.
The most dangerous vector in enterprise security is the browser, and it's the least hardened piece of software in the stack.
Harvest-now-decrypt-later makes post-quantum migration urgent today, not when the threat is realised — by then it's too late.
The companies that build AI products where the user owns the data will have the structural advantage in the regulatory decade ahead.
If this resonates, come work on it.
We're hiring senior browser engineers, a cryptography engineer, and a founding designer. We want people who care about the craft and the mission.