Engineering

Open-Source Contributions

There is a difference between a company that uses the Cosmos SDK and a company that wrote parts of it. The first reads the documentation. The second wrote the modules the documentation describes, found the bugs before they hit your mainnet, and stayed through the protobuf migration that touched nearly every line of the codebase. Vitwit is the second kind.

750+Merged PRs · Cosmos SDK
6Top-50 all-time contributors
7Team members across repos
9 yrsIn the Cosmos ecosystem
/ why it matters to you

The team that wrote it
debugs it differently

When you build on a chain, you are inheriting every design decision and every latent bug in its core modules. The team that wrote those modules knows the keeper internals, the migration paths, the parameter footguns, and the consensus-critical edges — because they argued about them in the original PRs. That knowledge does not come from reading a README. It comes from having shipped the code and lived with the consequences.

When a chain upgrade breaks something, or a module behaves in a way the docs don't explain, the team that authored that module isn't filing a support ticket with someone else. They are reading code they wrote. This is the practical case for working with upstream contributors rather than integrators.

/ what we built

Modules now standard
on nearly every Cosmos chain

Our development work centered on modules that ship with the SDK and run on every chain that builds on it.

01

Authz & Feegrant

Authorization and fee-granting primitives that let accounts delegate the right to execute messages and pay fees on behalf of others. These underpin gasless transactions, managed accounts, and a huge range of UX patterns.

  • Message delegation
  • Fee sponsorship
  • Foundation for DAO tooling
02

Groups

On-chain multisig and policy-based group accounts — the building block for DAO-style decision-making at the protocol level. Standard on nearly every Cosmos chain today.

  • Protocol-level multisig
  • Policy-based accounts
  • DAO governance primitives
03

Upgrade module

The module that coordinates chain upgrades. If you've ever watched a Cosmos chain hard-fork cleanly at a planned height, this is the machinery underneath. Getting it wrong halts the network.

  • Coordinated hard-forks
  • Planned-height upgrades
  • Consensus-critical path
04

Protocol pool, Server/v2 & Accounts

Newer architecture work on fee distribution, the next-generation server framework, and the accounts abstraction — the parts shaping where the SDK is heading.

  • Fee distribution redesign
  • Next-gen server framework
  • Account abstraction
The unglamorous, consensus-sensitive work nobody talks about
Protobuf migration — touched nearly every line of the codebaseConsensus pubkey rotationAutoCLI — automatic CLI generation from protobuf definitionsPorting Simapp from Gaia into the SDKFiltered pagination across query endpointsbech32 address migration for gRPC and IBC compatibility
/ research & proposals

Proposing where the protocol goes next

Contribution isn't only code — it is also proposing where the protocol should go. We've put forward research and improvements that shape the next version of the ecosystem.

The same instinct shows up in our bug findings: network halt from a PowerReduction parameter change, non-determinism when starting a new chain, state-load failures during upgrades, and a token-invalidation bug in the Game of Zones Hub. Finding these bugs before they hit your mainnet is exactly the work that keeps the ecosystem running.

IBC-PA
IBC-Constrained Policy AuthorizationA design for policy-based authorization that spans IBC channels — enabling cross-chain permission delegation.
ZK-Client
ZK-Cosmos light clientA proposal for a zero-knowledge-based IBC light client, enabling trustless verification with dramatically lower on-chain cost.
SDK #22001
Staking rewards distributionA proposal to adjust how staking rewards are distributed in the SDK, improving fairness across validator sets.
Commits across the wider ecosystem
Go relayerAkash NetworkRegen LedgerPassage3DPolygon HeimdallCheqdPalomaEmeris tracelistenerEmeris rpcwatcher

750+ merged PRs and six top-50 contributors you can go read.

A lot of infrastructure providers can run a node. Far fewer wrote the module the node depends on, found the bug that would have halted it, and proposed the standard that the next version will adopt. When you work with Vitwit, the team operating your validators, relayers, and chains is the same team that has spent nine years inside the codebase that all of it runs on.