The Uber fare that made us rethink dashcam evidence
It was 23:47 on a Tuesday when the message came in. A driver in Manchester had accepted a ride to the airport, completed it, and the passenger contested the fare. Nothing unusual there. What made us sit up was his next line: "I have the recording. I just have no idea how to prove it wasn't edited."
When a video isn't enough
That driver's frustration stayed with me for weeks. He'd done exactly what we'd hoped early Hawk users would do. He'd turned his iPhone into a dashcam. He'd caught the whole journey. He had GPS data, timestamps, the lot. And still, when it came time to dispute the charge with Uber, he hit a wall.
"They told me videos can be faked," he wrote. "So why would they trust mine?"
I'd spent years building apps for MRVL. We'd shipped things that worked, things people used. But we'd never had to think like someone preparing for a dispute. We'd never had to ask: what would a court accept? What would an insurance adjuster actually trust?
That's when we knew we had a problem with Hawk, even though Hawk didn't exist yet. The thing we were building couldn't just record. It had to prove.
Building the evidence, not just the footage
Our first instinct was to slap a timestamp on the screen and call it done. We've all seen dashcam footage with those big numbers in the corner. It's readable. It's there. But readable isn't the same as trustworthy. A timestamp can be overlaid after the fact. A video file can be copied, altered, re-exported a dozen ways.
We spent months talking to people. Insurance brokers. Police officers who'd seen evidence cases collapse. A small-claims court clerk in London who'd watched two years of disputes hinge on whether a video was original or not. One barrister friend of mine put it plainly: "If you can't prove the footage hasn't been tampered with, you might as well not have it."
That's when we arrived at SHA-256 integrity hashing. Every single clip that Hawk records gets a cryptographic hash written to it the moment the recording ends. The hash is a mathematical fingerprint. Change one frame, and the hash changes. Try to swap in a different video file, and the hash won't match. It's the same technology that protects banking transactions and medical records.
We knew it wasn't glamorous. No one gets excited about cryptographic integrity. But it meant that when someone sent a clip to their insurer or into a dispute, it came with proof. Not a claim. Proof.
The first driver who used it to win
Six months after launch, we got another message. This one was different.
A Lyft driver had been in a minor bump in central London. Another vehicle claimed it was entirely her fault. She'd used Hawk. Recorded it. Exported the entire trip, one tap, as a ZIP file with the SHA-256 manifest built in. She sent it to her insurance broker.
"They called back in two hours," she wrote. "They said the hash validation meant they didn't need to send it to a third party. They just processed the claim."
That message went straight onto the wall of our office. Not because it was marketing gold. Because it meant the thing we'd built actually worked the way we intended. The integrity hash hadn't been a theoretical safeguard. It had shortened her claims process by days.
After that, we knew we were on the right track. But we also knew we'd only solved half the problem. Proof was one thing. Access to that proof, and the control over it, was another.
Why the Evidence Locker matters
If Hawk could create court-ready footage, we had to make sure only the person who recorded it could access it. The Evidence Locker exists because we got tired of imagining scenarios where someone's phone was stolen, and three months of sensitive dashcam footage was sitting there unencrypted, waiting to be pawned or rifled through.
It's biometric. Fingerprint or face. If your phone is unlocked but Hawk isn't, you're locked out of those clips. We call it fail-closed design. If the biometric fails, the evidence stays locked. Not locked until you try again. Locked full stop.
The rideshare drivers liked this more than anyone else. They're recording passengers. They're recording late nights and difficult situations. The last thing they need is sensitive footage sitting around. One driver told us she felt "protected" for the first time since she'd started. Not paranoid about recordings leaking. Just protected.
From one dispute to policy change
Here's what I didn't expect: insurance companies actually started asking about us. Not as a vendor pitch. As a feature request. One major UK broker reached out and said they were seeing Hawk exports in claim files. They wanted to know if we could make it easier for them to validate the hashes on their end. (We could, and now we do.)
One insurer told their underwriting team that video evidence with SHA-256 validation would reduce their dispute-resolution time. Within a month, they'd changed their policy. If you can show them a Hawk export with an intact hash, they process it faster.
That's not a Hawk feature. That's something that happened to Hawk because we built it the right way first. We didn't optimise for flashiness or for the app store charts. We optimised for trust.
When I look back at that Manchester driver's message, I think about what would have happened if we'd just built a basic video recorder. He'd still be frustrated. The footage would still be questioned. And we'd have built a nice app that solved nothing.
What's changed since then
The app has grown. We added a cabin camera mode for rideshare drivers who wanted to record inside as well as out. We added shift mode so drivers doing multi-trip days didn't have to manage 50 separate clips. We added one-tap police-report submission for iOS, so if something serious happened, you could get your evidence to the authorities without formatting it yourself.
But the core hasn't changed. Every clip still gets hashed. Every export still comes with proof. Biometric lock still gates access. We didn't bolt those things on because they were trendy. We built them because a frustrated driver in Manchester asked a question we couldn't answer at the time.
Most of my job now is saying no to feature requests that sound good in a meeting but don't actually serve evidence. We don't auto-upload everything to a cloud server that we control. That would make marketing easier. It would make retention metrics look better. It would also make the footage not actually yours. Pro users sync to their own iCloud if they want to. The device stays the source of truth.
When's the last time you saw a video and actually knew whether it was the original or not? That's the question that built Hawk.