Comparison
Instabase runs inside your own cloud tenant with tier-one bank deployments behind it — an answer to the residency objection that does not require shipping hardware. Around US$400M raised and a valuation near US$2B. The honest difference is what you are expected to build after you sign.
| Instabase | Densery | |
|---|---|---|
| Residency answer | Deployed in your own VPC, with customer-managed keys. Sufficient for most institutional policies. | The same private-cloud option, plus genuinely on-premise on a GPU appliance for institutions whose obligation is legal rather than a preference — including air-gapped. |
| What you receive | A platform and a marketplace of building blocks — extractors, parsers, prebuilt solutions — that your team composes into an application. | A working document process on your first document type, in under three weeks. Not a runtime to build on. |
| Who finishes the last mile | Typically your engineers or a partner, using their tooling. Powerful if you have that capacity. | Us. If the deliverable is a capability your team still has to complete, we have not sold you anything. |
| Reference base | Named tier-one financial institutions and large mortgage originators. Substantially stronger than ours in the US. | Six named production deployments across Vietnam, Japan and Indonesia. No US customer we can name yet. Every one of the six will take a reference call. |
| Evidence layer | Logging, review and governance features available on the platform. | Per-decision provenance with confidence routing, approval gates and a correction loop, generated as a by-product rather than configured. |
| Scale | Roughly 400 people with an enterprise support organisation. | Much smaller. That means senior people on your account and no second line to escalate to. Both of those are true at once. |
| Best fit buyer | An institution with engineering capacity that wants to own the application it builds. | An institution with an operations problem and no appetite to run a build programme to solve it. |
Statements about Instabase are drawn from their public website and published materials as at 15 August 2026 and are our reading of them, not their words. Product capability changes; check anything here that matters to your decision, and tell us if we have it wrong.
Written straight, because you will find this out anyway and it costs us less to say it now.
Narrower than the list on the left, deliberately.
If you have the engineers, Instabase is a serious platform and the in-VPC story answers most residency policies without hardware. We would not try to talk a well-staffed platform team out of it.
Our whole argument rests on a specific observation: 88% of agent pilots never reach production, and the named blockers are evaluation, governance and reliability — not model quality. Those are the failure modes of a capability that was bought as parts and never finished. We sell the finished process because that is the step where the value actually is.
The fair test is to ask both of us the same question: what exactly is live, in production, ninety days after signature — and who is accountable if it is not?
Both are ungated and neither asks for an email. Run your own volume through the calculator, then open a real file in the audit-trail explorer and click the fields that failed.
The next step
A scoping session is not a demo. Bring twenty real files, redacted if you need to. We take three numbers off you — annual volume, fully loaded cost per file today, and what happens when the output is wrong — and hand back a one-page value case in your own KPIs.
If the arithmetic says we are not a fit, we will tell you in the room rather than six weeks later.
If your data can go anywhere and your documents are already clean and digital, you do not need us. Use a hyperscaler document API and spend the money on something harder.
Densery