We use privacy-friendly analytics to understand how the site is used. You can opt out any time on our Do Not Sell or Share page. See our Privacy Policy.
Most advice about finding automation gigs stops at "make an Upwork profile." That's half a strategy, and the missing half costs you. Watch one board and you're seeing 40 to 52 percent of the market depending on which one you picked, competing with everyone who made the same call.
Index window: August 20 to September 3, 2026. Every figure below carries its own sample size.
Freelancer.com and Upwork carry 92 percent of AI automation postings between them, and three smaller boards split what is left.
| Platform | Postings | Share |
|---|---|---|
| Freelancer | 13,214 | 52.0% |
| Upwork | 10,252 | 40.4% |
| Indeed | 1,588 | 6.3% |
| Skool | 168 | 0.7% |
| 167 | 0.7% |
DevSnipe scrapes all five continuously and classifies each posting into an automation category. Two things jump out. Freelancer.com posts more AI automation work than Upwork does, which contradicts almost every "where to find clients" guide written in the last three years. And the long tail is genuinely long: LinkedIn and Skool together account for less than one and a half percent of postings.
Neither fact means what it looks like it means.
Freelancer.com's 52 percent share is real, and it is the single biggest source of AI automation postings in the index. If you're optimizing purely for how many opportunities you see per day, adding Freelancer to your rotation more than doubles your visible market.
The catch is what those postings tell you. Freelancer's own listings show an average-bid figure rather than a client-stated budget, which is visible on their site rather than something DevSnipe measures. An average bid describes what other freelancers will accept, not what the client planned to spend.
All 13,214 Freelancer records in DevSnipe's index carry a bare dollar amount with no unit captured alongside it, so none of them can be classified as hourly or per project. That's why DevSnipe's rate index excludes Freelancer entirely, and you can read the full rate methodology and the numbers it does support if you want the detail.
So treat Freelancer as a volume play. You'll see more postings and get a weaker pricing signal from each one, which means more qualifying work per lead.
Upwork's share drops to 40 percent overall, and then it inverts once you filter for a specific tool. The pattern holds wherever the client had to know what they were asking for.
| Category | Sample | Upwork | Freelancer |
|---|---|---|---|
| n8n automation | 126 | 90% | 9% |
| Make.com automation | 45 | 87% | 13% |
| Cold email automation | 1,513 | 74% | 15% |
| Workflow automation | 1,031 | 71% | 20% |
| AI agent development | 1,110 | 70% | 16% |
| Web scraping | 393 | 62% | 38% |
Indeed, LinkedIn and Skool split the remainder. In AI agent development, at 1,110 postings, that split is 10 percent Indeed, 4 percent LinkedIn and 1 percent Skool. Note the sample column: the n8n and Make rows rest on 126 and 45 postings, so treat those two splits as a strong signal rather than a settled law.
A client who writes "n8n" in a job title has already done research. They know the tool, they probably know roughly what the build involves, and they're not going to be shocked by your rate. A client who writes "need automation guy" has done none of that.
DevSnipe watches all five boards and scores every posting against your stack, so the n8n and Make jobs surface without you refreshing anything. See how multi-platform job alerts work.
If you'd rather spend Connects than attention, we wrote up how to budget Connects for AI agent jobs.
DevSnipe sends you AI automation jobs from Upwork, Skool, and more before most people see them.
Indeed's automation postings are mostly jobs rather than gigs. That framing is a characterization of the platform rather than something DevSnipe measures directly, but the measurable part supports it: of 1,166 Indeed postings carrying a pay figure, only 162 state an hourly rate. The rest do not state one.
That makes Indeed close to useless for anyone billing hourly on short projects and genuinely useful for anyone who wants a six-month engagement at a predictable number. It's also the second reason DevSnipe's rate index is scoped to Upwork alone, since a non-hourly figure and an hourly rate cannot go in the same average.
LinkedIn and Skool are 0.7 percent each, 167 and 168 postings. There's no way to spin that into a headline. They're small.
What makes them worth a background glance is structural rather than measured. DevSnipe does not track how many freelancers watch a channel or how many proposals a posting attracts, so treat any competition claim here as an argument rather than a number. The argument is simply that far fewer people are watching a 167-posting channel than a 10,252-posting one.
Skool also behaves differently from a job board, though this is a read on the platform rather than a measured claim. Postings there tend to be a community member describing a problem rather than a client writing a specification, which means the work arrives unscoped and the buyer is already warm. That's a good trade if you're comfortable defining the project yourself, and a bad one if you need a written brief to quote against.
Across 4,318 Upwork postings with a stated hourly rate, the median AI automation posting advertises $22 per hour, and the top 10 percent advertise $55 per hour or more. Fixed-price work has a median budget of $200 across 3,714 postings. Those are advertised budgets rather than earned income, and the gap between the median and the top decile is the whole story of this market. The rate index breaks it down by category.
A $22 median doesn't move because you apply to $22 jobs faster. It moves when you're bidding at the specialist end of Upwork. DevSnipe does not cross-tabulate rate decile against category, so treat that link as reasoning rather than something measured. Our breakdown of what to charge for agent development work turns that into an actual number for a proposal.
Pick one board and you've capped yourself at just over half the market. But the fix isn't "check five websites every morning," because that's unbillable time you won't keep doing past week two.
A workable split, based on the distribution above:
Still deciding which niche to aim all this at? Start with how to start AI automation freelancing, then come back and pick your boards. For which skills are actually being hired right now, see our 2026 AI agent development demand report.
Watching five boards by hand doesn't scale, which is why DevSnipe exists. Postings get matched against your skills and the ones worth a proposal land in your inbox. Set up a skill profile and stop picking one board to lose on.
AI automation freelance jobs get posted across five platforms that DevSnipe tracks, and Freelancer.com carries the most volume at 52 percent of 25,389 postings collected in the two weeks ending September 3, 2026. Upwork follows at 40 percent, Indeed at 6.3 percent, and LinkedIn and Skool at 0.7 percent each. That ranking reverses for specialist work: of the 126 postings naming n8n, 90 percent were on Upwork. Watching a single board means seeing at most 52 percent of the market.
Freelancer.com publishes more AI automation postings than Upwork, at 52 percent versus 40 percent of the 25,389 postings DevSnipe collected in the two weeks ending September 3, 2026. Upwork carries the better-qualified work, holding 90 percent of the 126 postings naming n8n, 74 percent of 1,513 cold email automation postings and 70 percent of 1,110 AI agent development postings. Freelancer's own listings also show an average-bid figure rather than a client budget, which is visible on their site rather than something DevSnipe measures, and which says more about what other freelancers will accept than about what the job is worth. Use Freelancer for volume and Upwork for clients who already know which tool they want.
n8n and Make.com freelance jobs are both concentrated on Upwork. Of the 126 postings naming n8n in DevSnipe's index over the two weeks ending September 3, 2026, 90 percent were on Upwork, 9 percent on Freelancer.com and 1 percent on LinkedIn. Make.com follows the same shape across a smaller sample of 45 postings, at 87 percent Upwork and 13 percent Freelancer. These are the two most lopsided splits DevSnipe tracks, against an overall index where Upwork is only 40 percent of postings, and both samples are small enough to read as a strong signal rather than a fixed rule.
You can browse the live listings on the n8n automation jobs page.
LinkedIn and Skool each accounted for 0.7 percent of the 25,389 AI automation postings DevSnipe collected in the two weeks ending September 3, 2026, which is 167 and 168 postings respectively. Neither is a volume channel. DevSnipe does not measure how many freelancers watch a platform or how many proposals a posting draws, so the case for these two is structural rather than measured: far fewer people watch a 167-posting channel than a 10,252-posting one. Skool requests, in DevSnipe's reading of them rather than by any measure, tend to be community members describing a problem rather than clients publishing a specification, so the work arrives unscoped and the buyer is already warm.
AI automation freelance work advertises a median of $22 per hour. That median comes from 4,318 Upwork postings with a stated hourly rate, collected by DevSnipe over the two weeks ending September 3, 2026, and the top 10 percent advertise $55 per hour or more. Fixed-price projects carry a median advertised budget of $200 across 3,714 postings in the same window. These are published budgets rather than negotiated outcomes, so the median reads as the market's opening position, and the full rate index breaks them down by category.
The figures cover Upwork alone because DevSnipe captures no unit at all on Freelancer's amounts and Indeed's pay figures are mostly not hourly, so neither can go into the same average.
DevSnipe tracked 459 workflow automation jobs posted in the last 30 days, making it the third most active freelance category on the platform. Here's what the data shows about demand, rates, and which skills actually get you hired.
Web scraping leads every automation category on DevSnipe with 789 projects posted in the last 30 days. This report breaks down current rates by project type, the skills that win high-value work, and where to find clients who pay at the top of the range.
DevSnipe job data across 5 platforms shows which AI automation skills clients actually pay for in 2026. Agent orchestration, reliable API integration, and production error handling top the list, while flashy demos and feature memorization don't move the budget.
Get matched to jobs across Upwork, Skool, and more. Free tier includes daily alerts with 3 full job details per day.