“Watch me fix this site”: the dev-log format that finds clients who never search for a developer

We have watched how trade-skill content behaves on TikTok, and development sits in a genuinely lucky spot: the work is a performance that happens on a screen, which means it records itself. A carpenter needs a camera rig; you need OBS and permission. And the audience for it is far bigger than developers — there is a well-established appetite on this platform for competence-watching, the same pull that makes restoration and repair videos travel, and “watch me find out why this page takes eight seconds” scratches it perfectly for people who could not write a line of JavaScript.

That breadth changes who you are talking to. On a portfolio site you address someone already shopping; here the feed hands you three strangers at once — the founder with a quietly broken site, the aspiring dev studying how you work, and the civilian who just likes watching problems die. The founder is the one who becomes revenue, and the craft is narrating so the founder stays: name the symptom in their language in the first second (“this checkout loses people before it loads”), show the hunt, land the fix with the real before/after timing on screen. The aspiring devs and civilians are not wasted reach — they are the shares and follows that teach the algorithm to keep distributing you — but the script is always written for the smart non-developer.

The same walls from your feed-based channels apply with extra force, because screen recording captures everything. Never record over a client’s live dashboard, their customer data or an unreleased build; demo the technique on your own project or a sanitized reproduction instead, and show client work only once it is launched and cleared. Keep every number on screen a genuine measurement — this audience includes developers who will frame-step your DevTools — and skip any “this trick will rank you #1” framing, because you do not control rankings and the claim is the fastest way to read as a grifter. Within those lines, consistency is the whole game: the algorithm feeds accounts that post steadily, which is why this channel pairs naturally with your Instagram plan — one batch of captures, two feeds.

Lead with the fix-it arc, because it is the format the platform was built to reward: a problem anyone can see, a hunt with momentum, a payoff with proof. “This site takes 8 seconds to load — watch me find out why” is a complete sixty-second story, and the ending writes itself when the real timing drops on screen. Source the broken things ethically: your own older projects are a goldmine of honest befores, launched-and-cleared client work covers the rest, and when neither fits, rebuild the problem in a sanitized demo rather than pointing a camera at someone’s live business. Calling out a named stranger’s site uninvited might get views, but it reads as a public ambush to every founder deciding whether you are safe to hire.

Then run the dev-log, the format with the most durable audience: “day 3 of building this booking site,” the feature that fought back, the deploy that worked on the second try. Compress a session into forty seconds of jump cuts, narrate the decisions rather than the syntax, and let the code-to-live arc play out across a week of clips — a serialized build gives people a reason to follow rather than just watch. Tech-explained clips fill the gaps and pull the widest audience: what actually happens when you type a URL, why every site nags you about cookies, what “the site is down” really means. Every one of those is a founder-trust deposit, because the developer who explains clearly is assumed, correctly or not, to build clearly.

Round the mix out with the human layer — the setup tour, the honest “what I’d charge for a site like this and why” (real numbers you would stand behind, no bait rates), a launched client project celebrated with the client’s blessing and the collaborators credited. Cadence is where developer accounts die on TikTok, usually in week three when a sprint lands, so plan the whole mix on one calendar — our social media scheduler guide covers the mechanics, and a dependable scheduler is what keeps the streak alive through delivery weeks.

Here is the specific trap for developers on TikTok: the audience that makes your clips travel is mostly people who will never buy a website, so raw views are an even weaker signal here than usual. A million people watching you fix a flexbox bug is genuinely useful — it feeds the algorithm and fills the follower count — but the business lives in a thin layer underneath: profile visits after a fix-it clip, taps on the portfolio link in your bio, follows that arrive on client-project videos rather than meme-adjacent ones, and the comments and DMs that sound like owners (“my checkout does exactly this — do you take projects?”). Read those first, and treat the view count as weather.

Completion rate is the craft metric worth obsessing over, because it tells you whether your narrative arc holds. A fix-it clip people abandon at the midpoint has a saggy investigation; one they watch twice has a payoff worth copying. Saves lean practical — aspiring devs filing your technique, founders filing your profile — and shares are your referral network in miniature: a developer sending your clip to a group chat is vouching for you where subcontracts move. Watch which topics pull owner-shaped commenters: if the “why your site loses customers” explainers draw founders while the editor-themes tour draws students, you have just learned which clip to make more of, whatever the view counts say. Answer technical comments generously — visible expertise is the point — but move “can you look at my site?” to a real intake link instead of debugging a stranger’s production in public.

Close the loop weekly in one place: tag which clips preceded portfolio taps and inquiries, rebalance next week’s queue toward them, and keep the streak unbroken through delivery crunches — that steady cadence through your busiest stretch is exactly what a TikTok post scheduler for web developers is for, and it is the behavior this platform’s distribution most reliably repays.

A workable TikTok week for a developer assumes the footage is captured during real work and cut in one batch — nobody is filming fresh content mid-sprint. We spread posts across the week because this algorithm pays for regularity, we keep the serialized build running so there is always a reason to follow, and every clip’s numbers come straight from the terminal, not from the caption writer.

Week 1: Rig the capture pipeline

The whole system runs on footage, so start by making recording free. Set up a one-hotkey screen recorder, create a captures folder with a dump-everything rule, and record your actual work for a week without judging it — the bug hunts, the builds, the deploys, the before-states of anything you are about to fix. Sort your material sources into the three buckets: your own projects (always usable), launched client work (usable once a written OK is in hand — send those requests now), and NDA or unreleased work (never, and keep it out of frame even incidentally — check what is in your other monitor, your tabs and your notification tray before recording). Set the profile up to convert: a bio that says what you build and for whom, the own-work-and-credit line, and a portfolio link with a real inquiry path. By Sunday you want an hour of raw footage and zero clips published — the library comes first.

Week 2: Cut the first batch and start the streak

One sitting, one week of content: cut two fix-it shorts, one civilian explainer, two dev-log episodes and a setup clip from the raw pile, write hooks that name the symptom in the first second, and schedule all of it across the week. Then protect the streak over everything else — this algorithm pays for regularity, and the account that posts five times then vanishes for a sprint never compounds. Watch completion rate from day one: where people bail is where your edit sags, and tightening the hunt section of a fix-it clip is usually the fix. Answer comments fast and generously — a sharp technical answer in your comments is a free second impression — but route “can you look at my site?” to the intake link rather than performing free surgery on a stranger’s production in a thread.

Week 3: Find your repeatable formats

Three weeks of data will show you two or three formats that are yours. Maybe your DevTools hunts hold attention to the end while your talking-head clips die at second four; maybe the civilians love your “what the internet is actually doing” explainers while founders arrive through the checkout-speed clips. Double down on what holds and template it — same hook grammar, same arc, same payoff placement — because a repeatable format is what turns content from a creative act into a production line your schedule can rely on. Then audit honesty under momentum, because this is the week corners get tempting: is every number on screen still a real, reproducible measurement? Is the “before” genuinely the before? Has anything NDA-adjacent crept into a frame or a filename? One frame-stepped fabrication costs this audience’s trust permanently, and developers do frame-step.

Week 4: Systematize around the sprint cycle

Make the machine official. Recording is now a reflex (hotkey, dump folder, no judging), cutting is a recurring Friday block, and the queue stays at least a week deep at all times so a delivery crunch never breaks the cadence the algorithm is rewarding you for. Keep a running hook list — every bug you fix, every question a client asks, every “wait, how does that work?” from a non-technical friend is a future clip — and keep the serialized build alive, because a project with episodes is a reason to follow rather than just watch. Review the funnel weekly: clips to profile visits to portfolio taps to inquiries, rebalancing toward whatever owners actually respond to. The finished state is a discovery engine running on footage your real work already produces, honest to the frame, clear of every NDA, and costing you one editing block a week — while the feed quietly introduces you to founders who never knew they were looking for a developer.