Shake VS Bugpacker

Bugpacker's answer:

Too new to say. Bugpacker launched in September 2026.

Bugpacker's answer:

It has no backend. Every other tool in this category records your session, uploads it to the vendor's servers, and gives the developer a link. Bugpacker assembles the package in the browser and saves it to your downloads folder, so the bug data never reaches anyone's server, including ours.

The second thing is that it analyses the capture rather than only collecting it. Findings carry a confidence level, separate an observation from a defect, and refuse to overclaim: a duplicate element id is reported as consistent with input reaching one copy while something else read another, with an explicit note about what is not observable after the fact.

Bugpacker's answer:

For most teams the honest answer is that they should not, and the free tiers of the cloud tools are good. Choose Bugpacker when a link on someone else's server is not an option: the bug happens on a page with customer data on it, a contract forbids third-party processors, staging is only reachable over a VPN, or the security review that would approve a cloud vendor is longer than the bug is worth.

There is a second reason that has nothing to do with policy. The output is a file you own. A ZIP attached to a ticket still opens in three years, when a link to a vendor's page may not, and it keeps working after you stop paying.

Bugpacker's answer:

Developers, QA testers and support engineers who file bugs regularly and are currently doing this by hand: screenshot, paste, describe from memory.

The sharper segment is the same people inside a constraint: healthcare, finance, government, contractors under NDA, anyone whose staging is internal. They are not choosing between us and Jam on features. They cannot use Jam at all, and have been reporting bugs with a cropped screenshot and a paragraph instead.

Bugpacker's answer:

It came from the bug reports I kept receiving. Not bad-faith ones, ordinary ones from people doing their best: "the checkout is broken", a screenshot of a whole browser window, and nothing else. Each one cost an afternoon of going back and forth to establish what was actually on screen.

The tools that fix this all solve it by uploading the session somewhere, which is a fine trade for most teams and an impossible one for some. So the design constraint was set first: no server, at all, ever. Everything else followed from it, including the parts that were harder that way.

The piece that took longest was not capturing data. It was getting the findings to say what they could support and stop, rather than confidently naming a cause the capture cannot actually prove.

Bugpacker's answer:

TypeScript, React and Vite, built as a Chrome extension on Manifest V3. The ZIP is assembled in the browser with fflate, and the capture uses the Chrome extension APIs directly: scripting, webNavigation, webRequest and the side panel. Licensing runs against Lemon Squeezy's public License API, which authenticates with the customer's own key, which is what allows the product to ship with no server of its own.