News & Research

Build vs. buy e-signatures: what in-house signing really costs

Written by

Build vs. buy e-signatures: what in-house signing really costs

Building legally binding e-signing into your own product involves cryptographic sealing, audit trails, identity checks, ESIGN, UETA and eIDAS compliance, and per-customer data separation. This post puts numbers on that work... and gives a straight answer on where Firma.dev fits.

Short version: buy, unless regulation forces you to control the trust chain yourself. That holds when signing is your product, and it holds at very high volume. The rest of this post shows the working.

What you're signing up for when you build

Here's a planning estimate for a production-grade, embeddable signing feature in a multi-tenant SaaS app. These are our rough numbers for one experienced team, and yours will move depending on how much you already have (an email pipeline, a PDF library, a tenancy model).

  • Cryptographic sealing (6 to 10 engineer-weeks). PAdES signatures, certificate authority, keys in an HSM or KMS, RFC 3161 time-stamping, long-term validation data.

  • Audit trail and completion certificate (3 to 5 engineer-weeks). Append-only event log, signer IP and user agent, consent capture, certificate PDF tied to the document hash.

  • Signer identity (2 to 4 engineer-weeks). Email link plus one-time codes, session rules, auth method recorded per signer.

  • Compliance flows (2 to 3 engineer-weeks, plus counsel). ESIGN consumer consent and withdrawal, eIDAS AES requirements, legal review.

  • Multi-tenant isolation (4 to 6 engineer-weeks). Per-customer workspaces, scoped API keys, scoped tokens for embedded UI, per-tenant branding and sending domains.

  • Editor and signing UI (8 to 12 engineer-weeks). Field placement, templates, embeddable editor, mobile signing.

  • Long tail (6 to 10 engineer-weeks). Deliverability, reminders, expirations, signing order, webhooks, localization, accessibility.

  • Total: 31 to 50 engineer-weeks, about 7 to 12 engineer-months.

The sealing row is the easiest one to leave off, because a lot of people don't know it exists. An e-signature holds up in a dispute because of a digital signature over the document: a hash of the final PDF, signed with a private key, embedded in the file with the certificate that identifies that key. Change one byte and verification fails, and anyone can check it with standard tools years later. Getting this right means picking a PAdES level (you want B-LTA so it survives certificate expiry), running a CA, protecting keys, and contracting a timestampnig authority. None of that shows up in a sprint demo.

The ongoing costs nobody budgets for

Maintenance is the line item that never closes:

  • Certificates and keys. Rotation, renewal, revocation handling, and a plan for what happens if a signing key is ever compromised.

  • Trust questions from customers. Someone will open a signed PDF in Adobe Acrobat, see no green checkmark, and file a ticket. You'll need an answer and a validation tool.

  • Regulation. eIDAS was amended in 2024, national rules shift, and your enterprise customers will send security questionnaires asking exactly how your signing stack works.

  • On-call. Signing failures land at the worst possible moment, usually at month-end when every contract is due at once.

  • Deliverability. Signing emails that hit spam are signing requests that never get signed. Each tenant's custom domain is one more SPF and DKIM setup to babysit.

A reasonable planning figure is 20% of one engineer, permanently, once the feature is live... and that assumes nobody on the team leaves and takes the PKI knowledge with them.

The math

Assume a senior engineer costs €100,000 a year fully loaded. Swap in your own figure, the shape of the result barely changes.

  • Build: 9 engineer-months (the middle of the range above) is about €75,000.

  • Maintain: 20% of an engineer is about €20,000 a year.

Firma.dev charges €0.049 (~5¢) per envelope, with no monthly minimum and no per-seat fees. At that price:

  • The yearly maintenance budget alone covers about 408,000 envelopes a year, roughly 34,000 a month.

  • The initial build covers about 1.5 million envelopes before you've written a line of maintenance code.

If you send fewer envelpoes than that, buying wins on cost alone. And even past that point, the comparison assumes your in-house version matches a dedicated platform feature for feature, which is a lot to ask of a first version.

When building is the right call

There's one case where building makes sense: regulation requires you to control the trust chain yourself. Two examples:

  • Defense and classified work. A contractor whose documents have to stay on an air-gapped network inside the Pentagon's own systems, with signatures that chain to the Department of Defense PKI and the CAC smart cards staff already carry. Those files can't go to any outside service.

  • Government-run PKI. An agency required by policy to sign with certificates from its own PKI and to keep every document on infrastructure it operates.

In both cases the certificate authority, key storage and hosting are set by policy, so a hosted API can't meet the requirement no matter how good it is.

Signing as your product, and high volume

Neither is a reason to build.

Signing is your product. Use Firma.dev anyway and mark it up. White labeling, custom sending domains and per-customer workspaces mean your customers only ever see your brand, and you set your own price on top of €0.049 (~5¢) per envelope. Your team spends its time on the parts customers pick you for, and the sealing, audit and compliance work stays with people who do it full time.

You send a lot of envelopes. High volume makes the case for buying stronger. The more you send, the more an outage costs you, and the harder it is to justify running PKI as a side project when it isn't a core competency. At that scale you'll also get a better deal through a provider than you will building it yourself.

What Firma.dev covers out of the box

Every item from the build list above, mapped to what ships today:

  • Cryptographic sealing. Every completed document gets a PAdES-B-LTA seal with an embedded RFC 3161 timestamp, the full certificate chain and revocation data, so it stays self-validating long term. Anyone can verify a signed document independently at app.firma.dev/validate-signature.

  • Audit trail. Every document view, field interaction, signature, decline and admin action is logged with a timestamp, actor identity and IP address, and rendered into a completion certificate. Each signer's row shows how they authenticated, and the Signing Request ID is printed on every page of the signed PDF.

  • Identity. Email link, plus optional email one-time codes (valid 10 minutes, 3 attempts), recorded on the certificate with the time the code was verified.

  • Compliance. Firma.dev is compliant with the ESIGN Act, UETA, eIDAS (SES and AES) and UK eIDAS, and legally binding in 55+ countries.

  • Multi-tenancy. One workspace per customer, each with its own templates, documents and envelope usage. Workspace-scoped API keys and template-scoped JWTs for the embedded editors keep tenants separate, and white labeling plus custom sending domains keep the experience on your brand.

  • EU hosting. All infrastructure runs in the EU (Paris and Stockholm), with no replication outside it.

  • The long tail. Signed webhooks, reminders, 15 signing languages, anchor tags and conditional fields, all through one API.

And the gaps, stated plainly so you can decide with the full picture:

  • Firma.dev provides SES and AES. It doesn't issue qualified electronic signatures.

  • Signer identity is email-based (link plus optional email OTP). There's no SMS verification or national eID integration.

  • The PAdES seal is issued by Firma.dev's own certificate authority, so Acrobat won't show its green checkmark. The seal is fully valid, and the public validator above confirms it without relying on any PDF viewer's trust store.

If one of those gaps is a hard requirement for your documents, check it before you integrate. For everyone else... the integration is a few API calls and an embedded editor, and you can have signing live in your product this week.

If you're still set on building

Fair enough. Some teams have good reasons. We're putting together a practical guide to building your own signing stack, covering PAdES sealing, audit design and tenant isolaiton in detail, so you can go in with eyes open.

In the meantime, the fastest way to see what a finished version looks like is to try one. Get started with Firma.dev for free, no credit card required.

  1. Heading

Background Image

Ready to add e-signatures to your application?

Get started for free. No credit card required. Pay only €0.049 per envelope when you're ready to go live.

Background Image

Ready to add e-signatures to your application?

Get started for free. No credit card required. Pay only €0.049 per envelope when you're ready to go live.

Background Image

Ready to add e-signatures to your application?

Get started for free. No credit card required. Pay only €0.049 per envelope when you're ready to go live.