VI·SOFT
Starknet · Zero-Knowledge · Confidential Compute · Cross-Chain

Engineering partner for teams that can't afford to wait

We're a five-person team — a lead and four engineers — with years spent shipping Starknet infrastructure: appchain sequencers, STARK proving pipelines, TEE-based settlement systems. Not a generalist shop picking up cryptography on your project's dime.

Every engagement starts with engineers who've already shipped in this exact stack. You get production-grade code, reviewed trust assumptions, and a team that ships on the timeline a fundraising round or mainnet deadline actually demands.

See how we work
CairoRustSolidityStarknetSTARKSP1Groth16AMD SEV-SNPSolanaCelestiaCairoRustSolidityStarknetSTARKSP1Groth16AMD SEV-SNPSolanaCelestia
§ 01

What makes Visoft the right build partner

Working with a specialized team isn't about hiring extra hands — it's about removing the translation layer between what you need and what gets built.

Speed without the ramp-up

Engineers are matched to your stack from day one — Cairo, Rust, STARK proving, TEE attestation — so the first two weeks are spent shipping, not explaining what a Merkle proof is.

Production-grade quality

Code ships with test coverage, audited-style review practices, and documentation a future auditor or new hire can actually follow.

Flexible engagement, no lock-in

Scale from one embedded engineer to a full team, month to month, without long-term contracts or scopes that stop making sense mid-research.

§ 02

How we take a project from idea to production

Every engagement — a proving pipeline, a TEE-based settlement system, a cross-chain verifier — follows the same four-stage process.

01

Scoping and technical discovery

We map the protocol, threat model, and integration points with your team, so the architecture is validated before it's expensive to change.

02

Architecture and proof-of-concept

We design the system — circuits, contracts, or verification flows — and validate the riskiest assumptions first with a working proof-of-concept.

03

Iterative development in sprints

Engineers build in short, reviewable cycles with your team looped in continuously — no month-long silence before a surprise delivery.

04

Delivery, handoff, and ongoing support

Code ships with documentation and test coverage your team can maintain independently, with optional support for audits, upgrades, or scaling.

§ 03

What we build

Four domains, one engineering standard.

Starknet L2/L3 infrastructure

Appchain sequencers, settlement orchestration, and cross-chain messaging for teams running dedicated Starknet-based rollups.

Zero-knowledge proving pipelines

STARK proving and recursive verification, plus SP1/Groth16-based proving paths, sized to real constraints around prover time and onchain verification cost.

Confidential compute & TEE systems

AMD SEV-SNP attestation-based architectures for teams that need a hardware-backed trust anchor instead of a trusted operator.

Cross-chain settlement & data availability

Proof verification across chains and data-availability integrations for rollups that settle or publish outside their home chain.

§ 04

Engagement models built for startups and research teams

  • Embedded engineers who join your existing team and report into your process, not a separate agency workflow.
  • Fixed-scope builds for teams that need a defined deliverable — a circuit, an integration, an audit-prep pass — on a fixed timeline.
  • Flexible month-to-month scaling for research projects and startups whose direction shifts as findings come in.
§ 05

Why teams choose Visoft

Domain fluency, not just availability

Engineers arrive already fluent in the primitives — STARK proofs, TEE attestation, slot-ownership verification — so discussions move at the pace your team expects.

Built for scrutiny

Every deliverable is written assuming it will eventually face an external audit, a research reviewer, or a skeptical technical co-founder.

Continuity across the lifecycle

The same engineers who scope the architecture ship the code and support it post-launch — context never gets lost in a handoff.

§ 06

The expertise behind every build

STARK proving pipelines

Practical experience across STARK proof generation, splitting, and onchain verification, applied to real constraints around prover time and verification cost.

CairoSTARKStarknet OS

Confidential compute (TEE)

Hardware-attested execution environments used as a trust anchor for off-chain computation that still needs to settle onchain.

AMD SEV-SNPSP1Groth16

Cross-chain settlement & data availability

Verifying proofs across chains and integrating data-availability layers for rollups that settle or publish beyond their home chain.

SolanaCelestiaDA

Sequencer & appchain engineering

Sequencer and settlement-contract development for dedicated Starknet-based appchains, written with production operation in mind.

RustCairoSolidity
§ 07

Browse by category

Blog posts grouped by the systems and primitives they cover.

§ 08

Explore more

§ 09

What to know before you engage

Working with a specialized partner still means a few things are on you to get right.

IP AND OWNERSHIP ARE YOURS

All code, circuits, and documentation produced under an engagement are owned by the client on delivery — Visoft retains no rights to project-specific work.

SCOPING DETERMINES THE TIMELINE

Fixed-scope estimates are only as accurate as the discovery phase; skipping it to save time typically costs more time later.

NDAS ARE STANDARD, NOT AN UPSELL

Confidentiality agreements are signed before any technical discovery call, since most clients work on unannounced protocols or unpublished research.

§ 10

Frequently asked questions

Do you work with early-stage startups that don't have a technical cofounder yet?

Yes. We commonly work as the technical execution arm for non-technical founders, though we recommend at least one technical stakeholder on the client side to own long-term maintenance decisions.

Who owns the code and IP after the project ships?

The client does, fully, on delivery. Visoft does not retain licensing rights, equity stakes, or reuse rights over project-specific code or circuits.

Can you work with an existing in-house team instead of replacing it?

Yes — embedded engagements are designed to plug into your existing team and workflow rather than operate as a separate silo.

How do you handle confidentiality for unpublished research or unannounced protocols?

An NDA is signed before any technical discovery call. Research-stage clients can request that scoping discussions happen without any public-facing artifacts being created.

What if our technical direction changes mid-project?

Flexible and fixed-scope engagements both include a defined re-scoping checkpoint, so a pivot doesn't require canceling and restarting.