Winter 2022

Velt

SDK for adding an agent and human governance layer to your product

Visit website

What they do

SDK adding human and agent review and approvals to apps

How close to us

3, close

review and approval layer for agent output; checks mostly, sold to developers

Wedges

  • W2 Seeing whether your AI agents actually produce adjacent
L1 Company brain (knows the organisation) 0, none
L2 Agent team (runs agents, tools, approvals) 2, clear
L3 Does and learns (finishes, checks, measures, updates the brain) 1, light

Who buys

Firm size
not known
Buyer type
developer
Role
product engineer
UK
not known
Their own description (from YC)

For most of software's history the hard part was making things. Writing the code, drafting the document, putting the report together. That is the part that is getting cheap. An agent can now produce in an afternoon what used to take a team a month.

What doesn't get cheaper is verifying it. Someone still has to look at what the machine did, decide whether it's right, and answer for that decision later. Output goes up, review capacity doesn't, and that is where the work now piles up.

Velt is the infrastructure for that step, and it treats agents and people as the same kind of reviewer. A review agent takes the first pass and catches what a machine can catch. A person makes the call on what actually matters. Both work on the same object, in the same thread, under the same rules. Only after approval is the change applied, through your own webhook, and every decision is recorded, so you can always show who allowed what and when.

Teams build this themselves and it takes about two quarters. Comments, approvals, review agents, notifications, audit trails: the boring parts nobody thinks about until they're three months in. With Velt it's a few components you drop into a React, Next, Vue, or Angular app.

Full teardown

No full teardown yet. On the deep list.

Technical detail
Slug
velt
Lens hash
179096c479
Pass
screen
Stale
no