@timatbyteful

Founder @ Byteful | ex sneaker twitter

 Manchester, UK
byteful.com
Joined August 2024
  • Tweets239
  • Following355
  • Followers87
52Media
This week Max taught the whole company how 100+ servers talk to each other without the public internet ever finding out. Wisdom Wednesday is where one person at @byteful teaches everyone else something they know well. Max's was @Tailscale. One private network across datacentres, ISPs and cloud providers, nothing left open to the public web, and when a traffic rule changes you change it once instead of touching every machine. Open source and written in Go, which visibly pleased Max more than the security did. (We are still hiring a full-time Support Engineer to join us in person in Manchester. There will be more useful nerdiness.)
Max stands beside a whiteboard covered in hand-drawn network diagrams, explaining it to a seated colleague in Byteful's Manchester office during a Wisdom Wednesday session.

ALT Max stands beside a whiteboard covered in hand-drawn network diagrams, explaining it to a seated colleague in Byteful's Manchester office during a Wisdom Wednesday session.

1
5
470
Formatting a proxy list by hand is the sort of small job that eats twenty minutes and teaches you nothing... one missing colon and the import fails. Free formatter, whole list at once. It won't tell you which proxies are live, just how the list is written.
Byteful
@byteful15h
One missing colon can stop a proxy list importing 🔧 Our proxy formatter converts a whole list at once. Paste or upload, it works out the source format on its own, you pick the output, then copy or export. It rewrites the list. It will not tell you which proxies are live.
Byteful free tool. One missing colon can stop a proxy list importing. The proxy formatter converts a mixed list to socks5h in one go, without testing it. byteful.com/proxy-formatter

ALT Byteful free tool. One missing colon can stop a proxy list importing. The proxy formatter converts a mixed list to socks5h in one go, without testing it. byteful.com/proxy-formatter

16
Free data, long term discounting and direct access to our team. What more can you ask for.
Byteful
@bytefulOct 20
The Byteful Startup Program is live🚀 Free residential data for three months, then up to 50% off for two years. Three tiers for bootstrapped, VC or accelerator backed, and Y Combinator companies. Plus a Slack channel with a senior Byteful developer.
Free data for startups. Three months on residential, then founder pricing. A Byteful Startup Program banner shows three tracks: bootstrapped, VC or accelerator, and Y Combinator.

ALT Free data for startups. Three months on residential, then founder pricing. A Byteful Startup Program banner shows three tracks: bootstrapped, VC or accelerator, and Y Combinator.

1
2
44
The jump from 100 requests to 100,000 breaks things that looked fine in testing, and it's usually the boring stuff... timeouts set off the average, retries with no backoff, no real idea what your concurrency ceiling is.
Byteful
@bytefulOct 14
A scraping job that works at 100 requests does not automatically work at 100,000 📋 1️⃣ Data consumption at volume 2️⃣ Your timeout 3️⃣ Your concurrency ceiling 4️⃣ Retry strategy, with backoff 5️⃣ Whether you can tell why something failed
Five things worth checking before you scale: data consumption, timeout, concurrency ceiling, retry strategy, and knowing what broke.

ALT Five things worth checking before you scale: data consumption, timeout, concurrency ceiling, retry strategy, and knowing what broke.

2
43
Your browser can claim whatever it likes, but the TCP handshake underneath tells a different story... proxy servers run Linux. We rebuilt the stack in user space so you can set that layer yourself. Beta for now, so tell us if it does something odd.
Byteful
@bytefulOct 12
Introducing TCP/IP Fingerprint Emulation, now live in beta on datacenter and ISP proxies! Your browser says Windows. Does the TCP handshake agree? 💻 Windows, macOS, Linux, Android or iOS, set per request in your proxy username or once when you export your list.
1
58
We brought TCP/IP into user space?! Samsung smart fridge fingerprint please🤝
Byteful
@bytefulOct 8
Day two of @zytedata's Extract Summit is underway, and Byteful is here in Austin, Texas! Today at 12:00pm, our founding engineer Joe Dye takes the stage at Brazos Hall for a talk on TCP/IP fingerprint emulation🫆 Come and say hello before or after👋
3
130
First time in a driverless taxi > more comfortable / leg room > smooth ride > $7 compared to 15 dollar + tip @elonmusk bring to the UK
1
3
107
Two bot-traffic numbers came out this month and they do not agree. DataDome says bots and AI agents make up 26.5% of traffic across its customer base. Cloudflare puts automated traffic at roughly 57% of web requests. My guess is both are low. The agentic era is real and exponential.
Cloudflare
@CloudflareSep 30
More than half the traffic reaching sites on Cloudflare is now automated, and AI agents are the fastest-growing part of it. We're giving site owners the tools to see who's visiting, decide who gets in, and charge for access. cfl.re/4ynW3sS #BirthdayWeek

The Internet has a second audience

More than half the traffic reaching sites on Cloudflare is now automated, and AI agents are the fastest-growing part of it. We're giving site owners the tools to see who's visiting, decide who gets...

blog.cloudflare.com
56
A very cool example of what web-scraped data can enable in academic research. Found more than 70% of the thousand most repeated questions were templates or boilerplate. FAQ copy, repeated page elements, the same question published over and over. You can collect an enormous amount of the web quickly. Working out what the data actually matters is the tough bit. In the agentic age, that is only getting harder…😅
Common Crawl Foundation
@CommonCrawlSep 15
Bodhium Labs: From web pages to a question corpus FineQuestions draws on 110 snapshots of FineWeb, the filtered English-language corpus derived from Common Crawl, spanning 2013–2025. Extraction yielded roughly 13.4 billion raw questions filling 22 TB. Normalization and aggregation reduced these to approximately 1.5 billion unique questions.
2
41
sharing one login around a team was never a good answer, and it got worse once agents started holding the keys... organizations gives you roles and keys scoped to specific actions with an expiry. nothing to migrate, no extra charge.
Byteful
@bytefulOct 1
Byteful accounts now support organizations 👥 A shared workspace for your team with its own proxies, billing and API keys. Each member gets a role, so no more sharing one password. API keys can be scoped to specific actions with an expiry date, which is what you want when an agent holds the key.
Byteful Organizations launch. Shared team workspaces with member roles and scoped API keys, shown in the dashboard workspace switcher.

ALT Byteful Organizations launch. Shared team workspaces with member roles and scoped API keys, shown in the dashboard workspace switcher.

Creating an organization gives a team its own proxies, billing and members, while your personal account stays independent.

ALT Creating an organization gives a team its own proxies, billing and members, while your personal account stays independent.

API keys can be unrestricted with full access, or fine-grained with selected permissions and an expiration date.

ALT API keys can be unrestricted with full access, or fine-grained with selected permissions and an expiration date.

Existing accounts are already personal workspaces, with billing, proxy users and API keys intact. Nothing to migrate.

ALT Existing accounts are already personal workspaces, with billing, proxy users and API keys intact. Nothing to migrate.

3
59
selling concurrency in tiers always struck me as charging for the wrong thing... you pay for data or for ips, not for how many requests run at once. unlimited doesn't mean no safeguards though, and pretending otherwise would be dishonest.
Byteful
@bytefulSep 30
Unlimited concurrent sessions means there is no concurrency tier to buy 🔓 No plan-level cap to hit, no upgrade to reach for. You pay for data or for IPs, not for how many requests run at once. Fair usage safeguards still exist for traffic that threatens stability. Rarely triggered.
Cover slide on unlimited concurrent sessions at Byteful, and what unlimited actually means.

ALT Cover slide on unlimited concurrent sessions at Byteful, and what unlimited actually means.

Byteful sells no concurrency tier on residential, ISP or datacenter. You pay for data or for IPs, not for how many requests run at once.

ALT Byteful sells no concurrency tier on residential, ISP or datacenter. You pay for data or for IPs, not for how many requests run at once.

Unlimited concurrency means no tier and no plan-level cap. It does not mean no safeguards: fair use protections cover traffic that threatens network stability.

ALT Unlimited concurrency means no tier and no plan-level cap. It does not mean no safeguards: fair use protections cover traffic that threatens network stability.

Unlimited concurrency across residential, ISP and datacenter, with fair use safeguards for genuine abuse rather than normal scaling.

ALT Unlimited concurrency across residential, ISP and datacenter, with fair use safeguards for genuine abuse rather than normal scaling.

2
78
🔥
Noah Pham@itsNoahPhamSep 28
𝗘𝗺𝗗𝗮𝘀𝗵 𝟭.𝟬 𝗶𝘀 𝗵𝗲𝗿𝗲: a stable, open source modern successor to WordPress, built on Astro, with a secure, decentralized plugin registry! i’m incredibly proud to have contributed to this release during my internship at Cloudflare ⛅️ Read more: blog.cloudflare.com/emdash-c…
1
1
98
a page makes far more requests than the one you asked for, and most don't need a residential ip... paying residential rates for every asset a browser pulls is waste. smartpath decides per request, under 10ms. our own test last july: 40.79% less residential data.
Byteful
@bytefulSep 24
Smartpath decides in under 10ms whether a request actually needs a residential IP. If it does, it uses residential. If it does not, the request can go through datacenter instead without consuming residential data. Most request latency comes after that routing decision.
Smartpath saved over 40% of residential data consumption on average in a Byteful test across 10,000 pages in July 2025.

ALT Smartpath saved over 40% of residential data consumption on average in a Byteful test across 10,000 pages in July 2025.

The test covered 10,000 real pages across Amazon, Reddit, Footlocker, ASOS, Target, Redfin, Trustpilot and Bloomberg.

ALT The test covered 10,000 real pages across Amazon, Reddit, Footlocker, ASOS, Target, Redfin, Trustpilot and Bloomberg.

Smartpath assesses each request in under 10ms. Requests needing residential stay on residential; the rest route through datacenter without consuming residential data.

ALT Smartpath assesses each request in under 10ms. Requests needing residential stay on residential; the rest route through datacenter without consuming residential data.

A page loads more than its main HTML request, the routing decision happens per request in under 10ms, and savings depend on what each page actually needs.

ALT A page loads more than its main HTML request, the routing decision happens per request in under 10ms, and savings depend on what each page actually needs.

1
53
A 1996 protocol still carries most of the proxy world. The detail people miss: SOCKS5 takes hostnames, not just IPs, so DNS resolves proxy-side, in the region doing the work. Old spec, still quietly correct.
Byteful
@bytefulSep 22
SOCKS5 has been standardised since 1996 and defines three commands 🔌 CONNECT for TCP. BIND when the client needs to accept a connection back. UDP ASSOCIATE for UDP datagrams. It can also pass a hostname to the proxy for proxy-side DNS resolution instead of resolving locally.
1
99
Last week's Wisdom Wednesday 🧙‍♂ came directly out of work Denver Whiteley had been doing to further improve @byteful's security. He gave a presentation on infrastructure hardening and firewalls after learning more about both through actual work. That is pretty much the ideal source for these sessions... someone runs into something, understands it, then teaches the rest of us. Denver used a simple building analogy. An IP address is the street address, ports are the doors on the building, and behind each door sits a different application or service. Some doors need to be reachable for legitimate traffic. Others may be left exposed unnecessarily, which gives attackers another place to look. A firewall guards the street by only allowing legitimate traffic through the doors. That is the pattern with these sessions. Someone does the work, learns something doing it, then teaches the rest of us.
Denver Whiteley presenting a Wisdom Wednesday session in the Byteful office, with a slide explaining what a port is and marking Transport as layer 4 of the OSI model.

ALT Denver Whiteley presenting a Wisdom Wednesday session in the Byteful office, with a slide explaining what a port is and marking Transport as layer 4 of the OSI model.

16
66
👀
Eric Glyman
@eglymanSep 15
UK finance teams have waited long enough. Ramp is now live in the UK. Sorry we’re late. ramp.com/uk
2
103
Two new faces at Byteful, and both are already well into the work. Alex Kent joins Growth Operations, working on GTM. One of his first ideas is already days from shipping. Deni Mintsaev joins Support, already building out how we handle tickets. He also turned up with the exact same ergonomic keyboard as Max, which is either a coincidence or the most niche onboarding perk we never offered. Great having both of them here.
Deni Mintsaev and Alex Kent standing side by side giving thumbs up beneath the neon Byteful sign in the Manchester office, on their first weeks with the team.

ALT Deni Mintsaev and Alex Kent standing side by side giving thumbs up beneath the neon Byteful sign in the Manchester office, on their first weeks with the team.

1
65
sites can work out which os is behind a connection, and a mismatch stands out. you can now pick windows, linux or android, with live counts so you can see what's actually there. beta for now.
Byteful
@bytefulSep 10
OS Targeting is live on residential and mobile proxies 💻 Route your traffic through a Windows, Linux or Android device. Add os_windows, os_linux or os_android to your username, or pick it in the dashboard. Live IP counts per OS, same as the location filters.
OS Targeting in beta. Append an os parameter to your proxy username to route through a Windows, Linux or Android device, on residential and mobile proxies.

ALT OS Targeting in beta. Append an os parameter to your proxy username to route through a Windows, Linux or Android device, on residential and mobile proxies.

99
Real ones know 🥲
1
1
5
1,135
The good ol days
Nicolas Brillante
@nickbruhmanSep 9

Sneaker Twitter: The Most Competitive Market Nobody Has Heard Of

Article · @nickbruhman
2
169
most proxy debugging is guesswork after the fact. a status code and a request id on every response takes a lot of that away. worth reading the https bit though, tls means we can't see the target's own code, so we tell you what we did see.
Byteful
@bytefulSep 9
Every response carries its own debugging trail 🔎 A Byteful status code for what happened, and a request ID you can search in the live activity viewer to pull up the request details. For HTTPS we can't surface the target's own status code. TLS keeps it encrypted, so you get the failure we observed.
Cover slide for Proxy 101 episode six, on what happens when a proxy request fails and the debugging trail it leaves behind.

ALT Cover slide for Proxy 101 episode six, on what happens when a proxy request fails and the debugging trail it leaves behind.

Every Byteful response carries two debugging headers: x-byteful-status-code for what happened, and x-byteful-request-id to find the request in Live Activity.

ALT Every Byteful response carries two debugging headers: x-byteful-status-code for what happened, and x-byteful-request-id to find the request in Live Activity.

Six failure buckets on a request path. Five point to a specific cause, from credentials through to the target refusing. The sixth is a broader server or network error.

ALT Six failure buckets on a request path. Five point to a specific cause, from credentials through to the target refusing. The sixth is a broader server or network error.

TLS keeps an HTTPS target's response encrypted, so Byteful reports the failure it observed while connecting or transmitting rather than the target's own HTTP status code.

ALT TLS keeps an HTTPS target's response encrypted, so Byteful reports the failure it observed while connecting or transmitting rather than the target's own HTTP status code.

89
Load more