Skip to content
Velvt logo

Velvt

Tests whether an autonomous agent stays inside its granted spending, delegation and tool permissions, and records evidence of each run and repair.

AgentsEvaluationFor developersEnterprise
velvt.ai ↗X
At a glanceswept 2026-10-12

not read yet

Use

For
teams deploying autonomous AI agents that spend money, delegate, or make external commitments
Price
Price not read
Activity
Actively maintainedsite updated 2 days ago
Company
Domain since 2026domain registered
AlternativesSet it up for my teamIs Velvt yours? Claim the card →
What people do with it
  • test an agent against its spending permissions before raising limits
  • check whether an agent respects delegation and tool-use boundaries under pressure
  • confirm a repaired agent holds the same boundary on retest
The companyeach fact with its source · read 2026-10-13
  • Domain since2026domain registered
How a program reaches itfound, not claimed · read 2026-10-13
  • MCP serverai.velvt/velvtverified in the MCP registry
  • Stateanswers, requires sign-inchecked 2026-10-13
  • Address to connectwww.velvt.ai/mcp

MCP server details →

Each line is something a program found when it went looking: the registry entry, the server's own answer, a spec that parses. A missing line means we did not find it, not that it does not exist.

Alternatives in Agentsby use, among tools still maintained

Figure and bar: visits/mo, estimate · year: online since

All Agents tools →

Questions this page answers
  • What is Velvt?

    Tests whether an autonomous agent stays inside its granted spending, delegation and tool permissions, and records evidence of each run and repair. It is listed in the Atlas under Agents.

  • Who is Velvt for?

    teams deploying autonomous AI agents that spend money, delegate, or make external commitments

  • Do people use Velvt?

    Little sign of use.

  • Is Velvt still maintained?

    Actively maintained: site updated 2 days ago.

  • What are alternatives to Velvt?

    In Agents, by use: OneCLI, Aqta, Tandem, DataNexus MCP, Overwing, Prefactor.

Not quite it? Describe the job

beta

Five tools from the Atlas that fit it, each judged against what you wrote.

How we read a tool

No votes, no reviews, no vendor claims. Every week a crawler reads each tool's own site and a few public registries, and the words on the card are bands over what it read.

  • Use: visits to the site (estimated), installs from npm and PyPI, presence in Chrome's usage report.
  • Activity: the newest release on GitHub, npm or PyPI; the newest dated page on the site; open roles on a public jobs board.
  • Price: the vendor's own pricing page, read with the date. A figure is printed only when it is on that page.
  • The line and the use cases are written by us from the homepage, one tool at a time, and refused when they repeat the vendor's marketing.

How AI assistants read each site — the reading vendors ask us about — is on each tool's own visibility page.