Head-to-head · updated 13 September 2026

DataShield vs Basis Theory: is a payments vault the right home for your PII?

Basis Theory is a good programmable payments vault. Their hero says it plainly: "Built to keep customers transacting." Is that a card number, in a checkout flow? They hold it. They keep you out of PCI scope. Network tokens get minted. Old cards get refreshed too. They route the charge to whichever processor is up. They also post a starting price. Almost nobody else in this market does.

We sell a different thing. DataShield tokenizes whole datasets of PII and PHI at ingest. Then it governs what your AI agents may do with them. Agents query tokenized data over MCP. Turning a token back into real data is a privileged step. It's audited too. Here's where the two really differ, sources included. Yes, Basis Theory wins some rows.

Get free Demo Center access Get an instant quote

The short version

Pick DataShield when

The sensitive data is PII or PHI spread across datasets, not a card in a checkout. We tokenize at ingest. We keep joins intact across tables. We also generalize quasi-identifiers. Dates become year, decade or age band. ZIPs become 3 or 4 digits. Phones, SSNs and emails get partial masking too. Each column gets a measured score. See the data layer.
An auditor will one day ask you to prove an access log wasn't edited. Our chain answers with math. Run the verifier.
You need to pull an agent's access mid-session. The very next tool call should fail. It shouldn't wait for a token to expire. How Auth does it.
The vault, the policy engine and the evidence must sit on your own infrastructure. You hold the keys. Architecture.

Pick Basis Theory when

Your problem is cards. PCI Level 1, an Attestation of Compliance and answers for most of SAQ D sit on their pricing page. We have no PCI story at all.
You want payment performance, not just storage. Think network tokens, account updater, BIN data, and routing across several processors. A failed charge retries somewhere else.
You need to keep raw card entry out of your own systems, with client-side form elements that post straight to their vault.
You're a small team that wants to start on a published $995 a month plan and be live this week.

Bottom line: Basis Theory keeps card data out of your scope. DataShield governs what an agent may do with PII. It keeps proof of that, too. Plenty of teams run both. That's fine. Are agents touching health or customer records? Start here.

Feature by feature: tokenization platform against programmable payments vault

Competitor cells describe what Basis Theory's public site and docs say. We use the date above. Did we get something wrong? Email support@myorg.ai. We'll correct it, and credit you.

What matters	DataShield	Basis Theory	Edge
Audit evidence	SHA-256 hash chain with Ed25519-signed checkpoints that are themselves chained. Verification names the failure: tampering, insertion, deletion or truncation. Try the verifier.	Logs are a plan feature. 24 hours of retention, on the Starter tier. Custom access at Enterprise. We found no published cryptographic tamper evidence.	◆
Agent authorization	Every checked tool call passes a scope ceiling. It also passes a declared access tier. It gets a re-check before it goes out, too. A pulled grant lands mid-session, not after.	Their agentic commerce page describes scoped agent access. It covers permission controls for payment credentials, too. That checks the card, not the agent's wider authority.	◆
Break-glass	Scoped, time-boxed emergency access for agents. It auto-revokes on its own. It can't be quietly deleted from the log.	Not described in their public docs.	◆
GDPR erasure	Crypto-shred of per-subject key material. Plus ISO 27560 consent receipts too. The audit chain still verifies after erasure.	Token deletion via API. We couldn't find a documented erasure step tied to key destruction. Nor one tied to audit history.	◆
Tokenization	Deterministic, join-preserving, vault-reversible tokens. Applied at ingest, across whole datasets. Plus quasi-identifier generalization. Dates become year, decade or age band. ZIPs become 3 or 4 digits. Phones, SSNs and emails get partial masking too. Each column gets a measured score. We privilege each detokenization. We log it too.	A mature token API, with aliasing, masks and expressions. Plus proxy detokenization too. Built for one credential in one transaction, not joins across tables.	◈
PII and PHI scope	First-class. Classification and generalization ship in the base platform, always.	"Additional compliance options for PII and PHI data" sit at the Enterprise tier. The Starter plan is a PCI setup, plain and simple.	◆
Payments and PCI	No PCI attestation. Not our market. We won't pretend otherwise.	PCI Level 1, AOC included, and responses for 95% of PCI SAQ D. Home turf.	◇
Payment performance tooling	None. We don't touch authorization rates.	Network tokens, account updater, BIN service. Smart routing and backup routing across processors, too.	◇
MCP and agents	Native MCP endpoints sit on Auth, Ontology and Lighthouse. MCP tool tokens carry scope ceilings, always. Per-call metering ties right back to the agent.	An active agentic commerce push, since mid-2026. It's aimed at agent-initiated payments. No MCP server turns up in their docs.	◆
Deployment	Self-hosted, in your own cloud or data center, your call. Or on a dedicated single-tenant server we run for you. Your keys.	SaaS only. The Starter plan is a US setup. Other regions need an Enterprise call, it seems.	◆
Maturity signals	Auth, Guardian and Lighthouse are all live now. Guardian and Lighthouse have run since April 2026. SOC 2 isn't certified yet, and we say so plainly.	Founded 2020. PCI Level 1, SOC 2 Type II, ISO 27001, HIPAA. Named customers, monthly changelogs too.	◇
Pricing	Published model and a scoped instant quote, no sales wall.	Starter runs $995 a month for 20,000 tokens. Scale carries no price. Enterprise is a call. But they show a real number, and we respect that.	◈

◆ DataShield leads◇ Basis Theory leads◈ comparable

We draw Basis Theory claims from basistheory.com. Also their pricing and security pages, and their blog. Last checked 13 September 2026. We link them below, not just a paraphrase from memory.

Three things you get here that you won't get from a payments vault

Proof that survives an audit

A log can be quietly edited. That's worth little to an examiner. Ours is a hash chain with signed checkpoints instead. The verifier says what went wrong. Not just that something did. Two rules care about that most: EU AI Act Article 12 and HIPAA §164.312(b). Try it in your browser, no signup.

Authority that can change mid-flight

A nurse informaticist resigns on a Tuesday. Her agent is halfway through a 40-minute cohort job, unaware. DataShield re-checks the next call against current access, right then. It fails closed. A vault token just keeps working until it expires. How Auth does it.

An erasure you can defend

GDPR says delete. Your auditor says keep the log. Crypto-shred settles it, right there. We destroy the subject's key material. The data becomes unreadable, for good. The chain still verifies, every time. See the diagram.

Where Basis Theory is genuinely stronger

Let's be fair to them. Basis Theory has real certs on the wall. PCI Level 1, SOC 2 Type II, ISO 27001, HIPAA. They ship monthly, too. Their token API is older than ours. It's more polished too, for single-credential work. Network tokens, account updater and BIN data are real revenue tools. We will never build those, not our lane. And they publish $995 a month on a public page. Most of this market hides behind a demo form instead. That takes some nerve. It earns the pricing row as a draw.

Here's the push-back, in two parts. Their own blog makes an argument. A chatbot with access to card data lands you in PCI scope. They say AI is making that a problem nobody talks about. We agree, fully. But scope is only the first question. The second: was the agent allowed to make the call at all? The third: can you prove the answer, six months later? A vault log settles neither one. Not with 24 hours of retention on the entry plan. On their agent payments post, they make a claim. Agent identity and cardholder identity aren't the same thing. That sentence is the whole reason we exist. They wrote it about someone else's problem.

Questions worth asking both of us

These are the questions we'd want answered if we were buying. Ask them on each vendor call, including ours.

Can you cryptographically prove an audit log entry wasn't deleted?

DataShield: yes, it can. Each record commits to the one right before it. We sign and chain each checkpoint. Verification tells deletion from truncation from tampering, each time. Run it on a sample chain at /verify. Basis Theory: their pricing page treats logs as a retention feature. Just 24 hours at Starter. We found no tamper-evidence mechanism published. Ask them to show one.

What happens to a revoked agent mid-session?

DataShield re-checks access on each checked tool call, always. A pulled grant bites on the next call, no delay. Basis Theory describes scoped agent access for payment credentials. It's on their agentic commerce page. That's about what an agent may spend, not its wider authority. Ask how fast a pulled permission stops an agent already running.

Is Basis Theory a fair choice for PII and PHI? Not just cards?

It can be, at a price. Their pricing page lists it. "additional compliance options for PII and PHI data" at the Enterprise tier. So the published $995 Starter plan is a PCI setup. It's not a PII one. Ask what the PII tier costs. In DataShield, PII and PHI classification ships in the base platform. It covers 129 field classes. Quasi-identifier generalization ships there too. Dates become year, decade or age band. ZIPs become 3 or 4 digits. We score each column for cardinality reduction.

Do your tokens still join across datasets?

Ours do. The same input yields the same token, always. A patient ID in a claims table matches the one in an encounters table. Analysts can work without ever seeing the raw value. Basis Theory supports deterministic aliasing, too. But they tune the product for a credential inside a transaction. If your use case is cross-table analytics, ask for a worked example.

How does GDPR erasure interact with the audit trail?

DataShield crypto-shreds the subject's key material. Then it issues an ISO 27560 consent receipt. Actor identities in the chain carry HMAC commitments, not names. The proof still checks out, long after the subject leaves. Basis Theory offers token deletion. The effect on audit history isn't described in public docs. Ask whether erasing a subject breaks the log.

Does DataShield have SOC 2 or PCI?

Neither. We won't imply otherwise, ever. Basis Theory has both, plus ISO 27001 and HIPAA. That's a real gap. What we offer instead: Auth is live now. It ships a public threat model and a verifier anyone can run. Guardian and Lighthouse have both been live since April 2026. Design-partner terms include source escrow. Details on the security page.

Basis Theory's hero: "Built to keep customers transacting." Products listed include Payment Vault, Network Tokens, Account Updater and Enhanced BIN Service. - basistheory.com, 13 Sep 2026
Starter plan $995/month, 20,000 tokens, 1MB payload, 24 hours of logs, US environment only. PII and PHI compliance options sit at Enterprise. So do 95% of PCI SAQ D responses. - basistheory.com/pricing, 13 Sep 2026
PCI Level 1, SOC 2 Type II, ISO 27001 and HIPAA certifications. No published detail on key custody, residency or erasure. - basistheory.com/security, 13 Sep 2026
"Agent identity and cardholder identity are not the same thing." Agents authorised by mandates with spending ceilings rather than per-transaction approval. - Basis Theory blog, 8 Sep 2026
"AI is making customer support conversational, and it's becoming a PCI compliance problem nobody is talking about." - Basis Theory blog, 6 Aug 2026
EDPB names pseudonymization as a GDPR risk mitigation for AI models. - EDPB Opinion 28/2024

Other head-to-heads

Direct

Get free Demo Center access

You've seen the proof

Ready for a number? Scope your deployment and we'll price it against your own economics.

Get your quote →