Skip to main content

Context Data

Context Data makes InterLock, an open-source gateway that sits between AI agents and the databases, files and APIs they use. Each agent gets its own key, and every request is checked against roles and policies, with personal data masked before the agent sees it. Risky writes such as updates and deletes wait for a person to approve them, and every request lands in an audit log.

Traction 9/10
New York, United States · HQ
Founded 20242200+ followers
  • AI Agents
  • Artificial Intelligence
  • Enterprise Software
  • Software Only
Updated today

Funding

Raised to date

$120KRaised to date based on public sources. This may differ from the amount the company actually raised and is based only on what is publicly available on the internet.

Funding rounds are not available yet.

Founders

2 founders

Jide O.

CEO

Julia Kingsley

COO

Product

Problem

AI agents that need real company data are usually handed a database password, which works like a master key: a login that reads one order can read every customer's email address and may change or delete rows. Teams stall when agents move from demo to production because that access is hard to sign off on.

Solution

InterLock is a proxy that sits between AI agents and the systems they use, including PostgreSQL and MySQL databases, files in Amazon S3, internal APIs, and tools like Slack and GitHub. Agents connect to InterLock instead of connecting directly, and InterLock holds the real credentials while each agent gets its own key. Every request is checked in order: the key identifies the agent, roles allow or deny actions per table and column, and policies narrow access further. Results are scanned for personal data such as email addresses and Social Security numbers, which are masked before the agent sees them. Writes are graded by risk: adding a row usually goes straight through, while updates, unbounded deletes and schema changes queue for a reviewer, with an optional Slack notification. Every request, allowed or refused, is recorded in an audit log that exports as CSV. InterLock speaks the Model Context Protocol (MCP), the PostgreSQL wire protocol and plain HTTP, so agents connect by changing their address and key, not their code.

Target Audience

Teams putting AI agents on production company data, such as engineering and data platform groups building support assistants and analytics agents.

Features

  • Give each agent its own identity and key, and revoke access from an admin console at any time.
  • Limit an agent to specific tables or columns, with anything not allowed refused.
  • Mask emails, government IDs and other personal data in results before they leave your network.
  • Hold updates, unbounded deletes and schema changes for a reviewer to approve or reject.
  • Export the full audit log, including denied requests and what was redacted, as CSV.
  • Run it on your own infrastructure with Docker Compose or Kubernetes.
This profile is AI-generated and may contain inaccuracies.