Two questions eat a lot of developer and operator time. Why is this request slow, and where exactly did the time go? And how do you move a large file over a shaky link without starting from zero every time the connection drops?
BitSpeed answers them with two independent Go tools. BitSpeed Inspect records HTTP requests, shows which measured interval took the time, and compares saved runs. BitSpeed Transfer moves a file in verified chunks, and after a crash or a kill it sends only what the receiver can’t prove it already has. Each works on its own. Neither needs an account, a hosted service or an AI model.
The Pitch in Four Numbers
- 50 milliseconds of extra first-byte wait on a homepage, found by Inspect in one case study.
- 180 milliseconds of slowdown traced to a first-byte wait in another.
- 0 bytes resent that the receiver can already prove it holds, after an interrupted transfer.
- 2 tools, Apache 2.0, with the manual, tests and recorded results in the repository.
BitSpeed Inspect
Inspect explains one HTTP exchange from the client side. It records trace events, redirects, the body bytes it saw and partial failures, all as JSON, and it keeps three things carefully apart: what it measured, what it calculated, and what it’s suggesting the numbers mean. That’s honest about what timing can and can’t tell you. In the project’s words, a first-byte wait is evidence about an interval, not proof that a database or server CPU caused the delay.
Saved runs can be compared, so you can see what changed between yesterday and today, or between two servers. Inspect can also fail a CI build when a request gets slower, which turns performance into something a pull request has to pass. It handles POST requests and warm connections as well as fresh GETs.
BitSpeed Transfer
Transfer splits a file into chunks and verifies each one. If the transfer is interrupted, by a crash, a kill or a dropped link, it resumes and sends only what the receiver can’t prove it already has. Progress counts only what the receiver has confirmed, so the number on screen is real. It checks the whole payload before publishing it at the destination, and a clean command frees the disk space used for stored chunks once delivery is done.
It runs on Linux and macOS local filesystems and on local NTFS volumes on Windows, with prepare, send, status, serve and clean commands. Support for whole folders and several chunks at once is on the way.
Who It’s For
- Developers and SREs chasing down slow requests and wanting evidence, not guesses.
- Teams that want performance gates on pull requests.
- Anyone moving large files over unreliable links: field offices, ships, remote sites, home connections.
- Operators who need proof that what arrived is exactly what was sent.
Limits
Inspect’s timings don’t name a root cause on their own, and it says so. Recordings can contain private details, so review them before sharing. Transfer’s filesystem sync tests don’t prove every hardware power-loss guarantee, and some platforms are cross-built rather than tested natively.
The Four-Question Screening
Every project featured here gets the same four questions, the way you’d screen a startup idea.
1. Where is the money in it?
Both tools are free under Apache 2.0, with no paid tier. Each touches a paying market. Performance monitoring is a big business, and Inspect’s CI gate and run comparisons could grow into a hosted dashboard for teams that track latency across releases. Reliable file transfer is another: media companies, research labs, engineering firms and anyone shipping big datasets pay for managed transfer today. A paid Transfer with a relay service, team management and transfer receipts is a natural step.
2. What real problem does it solve?
Guesswork. “The site feels slow” turns into a long hunt without data that separates network time from server time. And a big upload that dies at 90 percent and starts over wastes hours and patience. BitSpeed gives the first one evidence and the second one a resume button that can be trusted.
3. The technology in plain words
Inspect is a stopwatch with lap times for a web request: how long to find the server, to connect, to get the first word back, to receive the rest. Transfer is a moving company that numbers every box and checks it off on arrival. If the truck breaks down, the next truck only carries the boxes that weren’t checked off.
4. Moat and differentiation
There are many HTTP timing tools and many transfer tools. BitSpeed’s edge is discipline: Inspect separates measurement from interpretation and won’t claim a cause it can’t prove, and Transfer only counts bytes the receiver can verify. Both are small, local and dependency-light. The moat that can grow is trust, built on that honesty and on recorded test results, plus the CI integration that makes Inspect part of a team’s daily routine.
Why It Matters
Speed problems and failed transfers are everyday friction. Tools that show where the time actually went, and that never make you send the same byte twice, save small amounts of time that add up across a whole team.
Follow the Project
Find out which tool you need, read about Inspect and Transfer, start with the quick start, or get the code from the BitSpeed repository on GitHub. BitSpeed is listed with the other open-source projects on SoftDevStudio.com, and more featured projects are in the Projects section.