Sales Tax Software for SaaS in 2026: Real Prices
Sales tax on software is the line item nobody puts in the budget model. In the US, roughly 20 states tax SaaS in some form. Each one draws the line in a different place. Sell into Europe and VAT applies from your first sale to a consumer, with no threshold waiting for you to get big first. Billing tools built for physical goods get both of those wrong by default, which is why founders end up shopping for a tax engine six months after launch.
I pulled the published pricing pages for the five services solo founders ask about most, checked in November 2026. I rebuilt the cost math at three revenue levels. Below is what each one actually bills on, plus what the entry price leaves out. The traps that only surface after you have signed the contract get their own section at the end.
The 30-second answer
| Service | Billing model | Published entry price | Filings included | Fits |
|---|---|---|---|---|
| Anrok | Tax software, per market | $100 per market per month | Yes, per the Starter plan | B2B SaaS with someone doing finance |
| Stripe Tax | Tax software, percent of volume | 0.5% per transaction (Basic) | Complete only, from $90/month | SaaS already billing through Stripe |
| Kintsugi | Tax software, per action | $0 monitoring, $75 per filing or registration | Yes, paid per action | US and Canada only |
| Paddle | Merchant of record | 5% + 50¢ per transaction | Paddle files as the seller | Small SaaS with no tax appetite |
| Avalara | Tax software, per state plus modules | About $69 per state per month, reported | Yes, sold as separate products | Mid-market with an ERP to feed |
What SaaS tax actually asks of you
Three obligations get bundled under the word “compliance,” and separating them is most of the buying decision. The first is knowing where you have an obligation: economic nexus in a US state or a VAT registration abroad. A marketplace that already collects on your behalf counts too, and it is the one founders forget. The second is charging the right rate on each invoice, which is a calculation problem with a long tail of edge cases. The third is filing and remitting on a schedule that differs by jurisdiction, which is paperwork you either do or outsource.
Most tools are strong on the middle step and thin on the ends. A calculator that returns zero for a location it does not support will pass every test you throw at it while quietly under-collecting. That single behavior is worth more of your evaluation time than the pricing table, because the failure is invisible until an auditor finds it.
Anrok: priced per market, not per sale
Anrok’s Starter plan is published at $100 per market per month, and the plan is aimed at founders who want the whole job handled rather than a calculation API. A market in Anrok’s terms is a jurisdiction you are collecting in, not a customer, so the bill scales with how many tax authorities you have registered with. Five active jurisdictions runs $500 a month, or $6,000 a year, before any custom terms.
The pricing is predictable in a way percentage-based tools are not, and that counts for something when you are forecasting. It also cuts the other way at the bottom: two markets at $200 a month is real money for a product doing $3,000 in monthly revenue, and no amount of accuracy fixes that math. The per-market structure rewards companies that have already decided to sell internationally. It penalizes the ones still testing. Custom pricing takes over past a certain footprint, and the published page does not say where that line sits. You can check the current tiers at Anrok’s pricing page.
One practical note on how the per-market bill behaves in a year where nothing much happens. The fee does not fall when your volume drops, and it does not rise when a single market suddenly accounts for most of your revenue. That symmetry is the point of the model. A flat commitment per jurisdiction is easy to hand to an accountant and easy to put in a spreadsheet, which is why finance people tolerate it. What it cannot do is shrink. If you register in eight states during a launch push and then consolidate your customer base, you are still paying for eight until you formally deregister, which is itself a filing with its own fee attached. Founders who plan to expand and contract should price that exit as part of the entry.
Worth reading alongside it: our breakdown of what usage-based billing actually costs, since metered pricing changes which jurisdictions you owe in and how much you remit each period.
Stripe Tax: 0.5% of what you collect
Stripe Tax Basic charges 0.5% of transaction volume in jurisdictions where you are registered, with no monthly fee and no filing. On $30,000 of monthly revenue that is $150, and it rises in lockstep with growth, which is fine at the start and less fine once you cross six figures a month. Tax Complete is the other product: a subscription starting at $90 a month on a one-year contract, with registrations, calculations and filings included, priced in tiers that scale with your business.
The catch with Complete is scope. It files the transactions that had Stripe’s automatic tax switched on. Invoices you issued manually are outside the filing. So are sales from before you enabled it, and anything billed through another system, unless you import them. US filing runs through TaxJar and needs a US bank account attached. If your revenue is all in one Stripe account and you turned automatic tax on at the start, this is the cheapest path by a wide margin. If you have three payment sources and an accountant who sends invoices by email, it is not. Current rates are on Stripe’s tax pricing page.
Kintsugi: free monitoring, $75 per action
Kintsugi splits the work differently. Monitoring is free, so the exposure question costs you nothing to answer. Each filing or registration then runs $75 on the Starter plan for US and Canada. Scale and Premium move to flat monthly subscriptions with custom quotes, and there are no onboarding fees or transaction fees on any tier.
For a founder with one US state and no international sales, that is a good deal: one registration and twelve monthly filings is $975 in year one, which undercuts every subscription on this list. It gets expensive in proportion to how spread out you are. Ten jurisdictions filing monthly is 120 actions, or $9,000, and at that point a flat plan from a competitor usually wins. The other limit is geography. This is a US and Canada tool, so if you sell to European consumers you still need something else for VAT. Pricing is at Kintsugi’s pricing page.
Paddle: you stop being the seller
Paddle is not a tax tool, it is a merchant of record. It buys your product and sells it to your customer. It also carries the tax liability itself, at 5% plus 50¢ per checkout transaction with no monthly fee. That is roughly double a typical card processing rate, and the difference is what you pay to never think about registrations again.
For a founder at $5,000 a month who would otherwise owe $200 to a tax engine and a weekend of filing, Paddle is the rational choice. At $100,000 a month the 5% is $5,000, against perhaps $400 for a tax engine plus the card fees you would have paid anyway, and the arithmetic turns against it fast. There is a strategic cost too. Paddle is the seller of record. That means the customer relationship and the billing data live in their system, and migrating off later means moving subscribers. If you are weighing the two paths, our comparison of Stripe against Paddle and Lemon Squeezy goes through the fee stack line by line.
Avalara: the enterprise default
Avalara does not publish a simple price, and its US SST plan is reported at around $69 per state per month when billed annually, with filings and returns sold as separate products on top. It is built to sit behind an ERP and a finance team, and the implementation is a project rather than an afternoon. Nothing about it fits a solo founder at $2,000 MRR. It is the name your accountant will recognize without a search, though. That recognition matters when you are handing off compliance to someone else.
Cost math at three revenue levels
Assume a SaaS on Stripe, selling to US customers in five states, with modest European sales through a single OSS registration. Six jurisdictions in total. The figures below use published entry pricing and exclude card fees.
| Monthly revenue | Anrok (6 markets) | Stripe Tax Basic | Stripe Tax Complete | Kintsugi (72 actions/yr) |
|---|---|---|---|---|
| $3,000 | $600/mo | $15/mo, no filing | $90/mo | $450/mo equivalent |
| $30,000 | $600/mo | $150/mo, no filing | $90 to $200/mo | $450/mo equivalent |
| $100,000 | $600/mo | $500/mo, no filing | Tiered, quoted | $450/mo equivalent |
The shape of that table is the whole argument. Per-market and per-action pricing stay flat as you grow, while percentage pricing climbs. A founder planning to triple revenue in a year should treat a percentage fee as a variable cost with no ceiling. A founder unsure whether the product will survive the quarter should treat a $600 monthly commitment as the riskier line item.
One more data point worth having: GoodVat publishes flat monthly plans from $20 for 250 transactions up to $100 for 2,500, with ready-to-file returns included, which lands between the two models. It is a newer entrant without the integration list of the incumbents. Like any vendor’s own comparison, its published table should be read as a claim rather than a verdict.
The traps that show up after signup
Four things bite founders in the first year. Percentage-based calculators only tax transactions where you are registered, so the number looks small until you add a state. Merchant-of-record pricing hides the processing rate inside the headline number, which makes the comparison against Stripe feel better than it is. Filing services charge per jurisdiction, so a tool that looks cheap on the pricing page gets expensive the moment you have nexus in eight places. Every one of these tools also needs your product catalog mapped to a tax code, which is a task rather than a toggle.
There is also the question of what happens when a jurisdiction is not supported. A calculation that returns zero with a reason code is technically correct and practically dangerous. Before you commit, ask each vendor what their API does for an unsupported location. Check whether the answer is a refusal or a silent zero.
How I would choose
If everything you sell runs through one Stripe account and you enabled automatic tax from day one, take Stripe Tax Complete and stop reading. If you sell to businesses in a handful of countries and have someone who can own the filing calendar, Anrok’s per-market price buys accuracy plus a person who answers email. If you are US-only, Kintsugi’s free monitoring plus $75 per action is the cheapest honest option until you pass about six jurisdictions. If you want the liability to disappear, Paddle’s 5% is a real cost and a real service, and the break-even against doing it yourself sits somewhere around $15,000 a month.
Whichever you pick, decide before you cross the first threshold rather than after. Backfiling is the expensive version of this problem, and the tools that make it cheap later are the same ones that make it cheap now. If pricing your product is still open, start with how to price your first digital product so the tax line lands in the model instead of on top of it.
Disclosure: some links above are affiliate links. Prices were read from each vendor’s published page in November 2026 and can change without notice. Nothing here is tax advice; confirm obligations with a professional in your jurisdiction.
