Skip to content

Loophold

AI agents at work on your systems, within the rules you set

Loophold stands between AI agents and your company: you write the rules, the people responsible approve the sensitive actions and every step is recorded in a tamper-evident log. Agents never get hold of your credentials.

Who it’s for: Companies using AI agents on code, email and databases, and the system integrators who run it for their clients.

console.tuaazienda.it
The Loophold console: an AI agent’s email to an external recipient waiting for the owner’s decision, with the rule that stopped it, the evidence collected by the Gate and the buttons to approve or reject it

What Loophold offers

  • Rules in plain language

    Describe a rule as you would to a colleague and it becomes exact conditions you can read, per person, team or system. Each rule allows, asks or blocks.

  • People decide what matters

    Sensitive requests wait for the people responsible, on the web, Telegram or Discord. With no answer by the deadline, the request is rejected.

  • Credentials out of reach

    Keys stay in a vault and Loophold’s Gate uses them to act. An agent holds only a token that can ask.

  • Sensitive data stays hidden

    Databases are connected read-only, and rules mask IBANs, tax codes, cards, email addresses or whole columns in every answer.

  • Teams and access limits

    Teams and roles, a set of systems for each person that an agent can only narrow, and any agent can be suspended at any moment.

  • A tamper-evident log

    Each event carries the fingerprint of the one before, so any change shows up when the chain is verified.

A look at Loophold

Rules you can read before they go live

Write a sentence and Loophold turns it into conditions you can check and change: who it applies to, on which system, when and with which exceptions.

  • Allow, ask a person or block: when rules overlap, the strictest wins
  • It looks at what the action contains: files, recipients, tables
  • Tried on sample requests before activation
  • Every change is saved as a new version
console.tuaazienda.it
Writing a rule in Loophold: the sentence in plain language and the exact conditions the Gate will apply, each one editable

Who did what, on the record

Every request, rule, decision and execution joins a chain of fingerprints that can be verified even outside Loophold.

  • Who asked, which rule applied, who decided and what was executed
  • CSV evidence export for customer and auditor questionnaires
  • Full detail kept for 90 days by default, adjustable from 30 days to a year
console.tuaazienda.it
The Loophold log: the chain check, agents’ actions with their outcome and who decided, and the full history of one request from rule to decision

How it works

  1. 01

    The agent asks

    It holds a token that can only ask: every push, email or query reaches Loophold as a request.

  2. 02

    Rules and people decide

    Your rules apply to every request. When one requires it, the agent waits for the owner’s answer; everything else goes through on its own.

  3. 03

    The Gate acts and records

    The action runs with the key kept in the vault and goes on the log. A refused agent is told which rule stopped it and what to do instead.

Frequently asked questions

What is Loophold?

A control layer between AI agents and a company’s systems. Agents go through it before acting: it applies the rules, involves the people responsible when a rule calls for it, acts with credentials the agents never see and records each step in a tamper-evident log.

What can the rules control?

Who an action is for, the system, the time and the content: the branches and files of a push, the recipients of an email, the tables of a query. Each rule allows, asks a person or blocks, and the strictest prevails. An action covered by no rule goes through and is logged.

Which systems and agents does it work with?

Code repositories on any host, company mailboxes including certified email (PEC), and databases in read-only mode. Claude Code and Codex connect with one command; n8n, other AI assistants and custom agents with a Gate token.

Do agents ever see the credentials?

No. Credentials stay in a vault, on your server or your integrator’s, and the Gate uses them to act. Agents only hold a token that can ask.

Can an agent get around Loophold?

Only through access it still has elsewhere, because the Gate controls what passes through it. Each agent therefore comes with a checklist of what could still bypass the Gate and how to close it.

How is AI used?

Exact checks involve no model. AI translates rules written in plain language and, if you choose, judges a meaning, such as whether an email contains confidential data. It can stop a request or send it to a person, never let one through, and each judgment is logged.

Does it make us compliant with NIS2, DORA or GDPR?

Not on its own, as with any tool. Loophold provides the evidence: an export for customer and auditor questionnaires showing where each credential is held, what was decided and what was sent to an AI provider.

Does Loophold monitor employees?

No. By default the company sees totals and each person sees their own activity. Views of individual activity only turn on if the owner states a legal basis, and that choice is recorded.

Who runs it, and in what language?

A system integrator installs and runs Loophold for its client companies, on its own server or cloud and under its own brand, with separate rules, systems and log for each client. The console and rule writing are in Italian for now.

Information about Loophold

Write to us for a demo or a quote: we’ll show you Loophold in action and answer all your questions.

Chat with us on WhatsApp