Sidekickブラウザ: what the free tier covers

Pricing questions about Sidekick the browser now come with a complication. The product cannot be bought. It was sunset on August 3, 2025, and its official address returns a permanent redirect to a different company's product. The pricing structure is still worth reading, because the line it drew between free and paid sits in almost exactly the same place across every product still shipping in this category. Understanding where that line falls, and why, makes the next pricing page take two minutes instead of twenty.

What the free tier actually covered

Two plans were published. The free tier allowed 5 sidebar apps, 10 apps in total, and 2 saved sessions, plus ad and tracker blocking, a global search across open apps, the automatic tab suspender, and standard support.

Pro was 12 USD per month. It added multiple workspaces, described as beta, multiple accounts of the same service, split view, password sharing, custom web apps built from any URL, and the removal of the session and sidebar limits.

Read as a design decision rather than a price list, the split is coherent. Everything that tidies a messy tab strip was free. Everything that separates identities was paid. Someone evaluating the product to see whether it could calm a cluttered browser could do that at no cost. Someone evaluating it because a work Gmail and a personal Gmail keep logging each other out was looking at a paid feature from the first minute.

That is the pattern to carry into any comparison. A free tier in this category is best read for what it leaves out, because what it leaves out is usually the reason the search started.

Where a ten app allowance runs out

Ten apps sounds generous, and by the numbers it is at the wider end of the category. In practice the allowance tends to go unspent while the user still ends up on a paid plan, which sounds contradictory until the counting method changes.

Count the services that stay open all day and the list lands around six or seven. Mail, chat, documents, a tracker, a calendar, expenses, maybe a CRM. That fits inside ten with room to spare. Now count the identities instead. The company Gmail and the personal Gmail are two. The internal Slack and the client's Slack are two. The personal Notion and the team Notion are two. The service count has not moved, the identity count has roughly doubled, and identities were the paid column.

The two session limit had the same shape. Sessions were meant for folding away a research pile and bringing it back later. Anyone running three projects in parallel has three piles, not two, so the third gets rebuilt by hand every morning.

Neither limit was unreasonable. Both illustrate the same evaluation mistake, which is reading a free tier as a quantity rather than as a unit. The number only matters once the unit matches what actually needs separating.

What happened to money already paid

This is the part of a pricing question that rarely gets asked in advance and always gets asked afterwards.

Sidekick put out a message during 2024 indicating it would sunset by the end of that year. It then kept running and kept accepting Pro subscriptions into January 2025. The product had also been sold as a lifetime deal through AppSumo, and holders of those deals were openly unsure whether any shutdown was genuine, since the announced ending had already come and gone once. The actual stop arrived on August 3, 2025, more than half a year after the notice that people had stopped believing.

A monthly subscription limits exposure to whatever is left in the current billing period. A lifetime deal does not. Its whole value rests on the product continuing to run, and that condition never appears on the checkout page beside the discount.

The instinct to avoid lifetime deals is not quite the right conclusion either, because a monthly subscription also stops working the day the product does. The difference is the size of the loss rather than its existence. What actually reduces risk is looking at signals outside the pricing page: how often the app ships updates, when the documentation was last touched, how large the company is, and whether the product falls back to a free mode or refuses to launch when a subscription lapses. That last one is usually stated somewhere in the help pages and is worth finding before paying.

The category priced side by side

Based on what each vendor publishes on its own site as of late 2026, the free allowances and paid prices line up like this.

Product	Free tier	Paid
Sidekick (discontinued)	5 sidebar apps, 10 apps total, 2 sessions	Pro at 12 USD per month
Wavebox	2 spaces with 2 apps each, 1 extension, 1 dashboard	99 USD per year, roughly 8.33 per month, Teams at 12.50 per user per month
Shift	5 spaces with 10 apps each	Advanced at 199.99 USD per year
Rambox	Unlimited apps, no workspaces	Pro at 7 USD per month or 70 per year, teams at 14 per user per month with a 5 user minimum
Ferdium	No limits, Apache 2.0 open source	None

Comparing only the right hand column produces the wrong shortlist. The column that decides the outcome is the free tier, and specifically whether a second account of the same service fits inside it. In most of these products it does not. The pricing page sets out how that line can be drawn differently, and the comparison with Wavebox shows what the same money buys in a product with a broader feature surface.

Why the free line sits in different places

The allowances differ because the cost structures differ, and knowing which model a vendor is using makes an unfamiliar pricing page readable at a glance.

Some price by app count. Wavebox gates spaces and apps, on the logic that each resident window is load. Some price by identity count, which is what Sidekick did, on the logic that a single user consuming more accounts should contribute more. Some price by sync and sharing, because configuration stored in a vendor's cloud and passwords shared across a team carry direct server costs that are hard to give away. And some do not gate features at all, which is the open source position: Ferdium ships under Apache 2.0 and depends on volunteer time rather than revenue.

The practical consequence is that the cheapest option flips depending on the shape of the requirement. Someone needing fifteen apps with one account each and someone needing five apps with two accounts each will reach opposite conclusions from the same table.

One more thing about these lines: they move. Free allowances in this category have narrowed before, and a decision built entirely on a current free tier should include a plan for what happens if it changes.

Why a free trial often fails to answer the question

There is a specific trap in evaluating this category on a trial, and it follows directly from where the paid line sits.

A trial of a product whose free tier excludes multiple accounts gives an accurate picture of everything except the feature being evaluated. A week spent with a tidy sidebar and a fast global search feels good, confirms nothing about the actual requirement, and ends with a decision made on the wrong evidence. When a paid trial is available instead, the opposite risk appears: the trial period is usually short enough that the configuration never gets finished, so the thing under test is the setup process rather than the daily use.

A more reliable approach costs nothing. Build the identity list first, read it against the free tier column, and reduce the field on paper. Whatever survives is worth installing, and it is usually two candidates rather than six. Then run each one with the real accounts rather than a sample set, because the failure modes in this category show up specifically when two accounts of the same service are signed in at once, and a sample set of unrelated services will never reproduce them.

Set a fixed evaluation length before starting, because these products reward habit and punish sampling. Two or three days is long enough for the annoyances to surface and short enough that a failed candidate has not absorbed a week. What to check on the last day is simple: whether reaching for a given app still requires searching for it.

Which services can be installed as pinned apps also varies, and that is worth checking before the trial rather than during it. A list such as supported apps answers in seconds whether the specific tools in daily use are covered.

Three costs that never appear on a pricing page

Comparing monthly figures leaves out the expenses that actually decide the total.

• Migration. Bookmarks and passwords export cleanly from any Chromium based browser. The layout of containers and pinned apps does not, because no standard format exists for it. Rebuilding ten apps across several identity containers takes one to two hours even with practice, and it recurs every time a product is replaced.
• Account overhead. Several products in this category create a vendor account during installation. That is one more credential to manage, and on a company machine it can require checking an internal policy before the browser is usable at all.
• Continuity. Whether an installed copy keeps working after a vendor changes direction is determined by the licence, not the price. Published source means an existing build keeps running. A closed product stops being installable on a new machine the day distribution ends. The comparison with Ferdium covers what that guarantee costs in polish and support.

Adding those three to an annual figure changes the ranking more often than the subscription price does. A product that is 30 USD per year cheaper but forces a rebuild after eighteen months is not cheaper for anyone whose hourly rate is worth counting. Conversely, a slightly more expensive product that exports its configuration and has shipped updates steadily for years is buying something the price column does not describe.

Sidekick supplies the worked example for the third one. Twelve USD per month was a fair price for what it did, and it was a price that only meant anything while the product was running.

What to change first

List every identity that has to stay signed in simultaneously, counting email addresses rather than service names, and put that number against the free tier column of each candidate. Convert whatever remains to an annual figure so monthly and yearly prices can be compared honestly, then add the migration hours at a realistic rate. Most of the time the gap between 7 and 12 USD per month stops being the deciding factor, and where the free line falls becomes the whole decision, including on the SpaceDeck page.