Early accessTokenalytics Factory is opening to a small group of builders

Security

Your accounts, your machine, your merge

Tokenalytics is designed so agents can do useful work without holding more access than a stage needs.

This page describes the early-access design. A full security document will be published before general availability.

Principles

Model providers

Code and context that agents need are sent to Anthropic or OpenAI under your account and their terms. Choose providers and accounts that match your organization's policies.

Feedback screening

Feedback cards are screened for spam and personal data before an agent reads them. Screening reduces risk; it does not replace your own data-handling rules.

Reporting a vulnerability

Email dev@abld.site with steps to reproduce. Please do not disclose the issue publicly until we reply.