ATP PETRICHOR
The enterprise AI gateway
One key · Every model — the rain that awakens intelligence
Scaling AI adoption doesn't come from more keys — it comes from a governance system. ATP Petrichor lets an enterprise manage AI by organization: who can use which models, how much each department spends, and what happened in every request, all in a single console.
GOVERNANCE
Four layers that turn AI usage into a tree you can see
Organizational governance
A four-layer structure of organization → workspace → project → key, with quotas allocated down the hierarchy and available, allocated, and consumed amounts visible at every level.
Project-level model authorization
Set the allowed model list at the project level, so every key in that project can only call authorized models — permissions enforced, not left to good behavior.
Quota & budget control
A quota-based budget model: departments get alerts before they exceed limits, contract quotas and top-ups run in parallel, and AI spend turns from an end-of-month surprise into a live dashboard.
Request logs & auditing
Every request records the model, key, token usage, and status; keys stay in the audit log even after revocation — the evidence a compliance review needs is built in.
HOW IT WORKS
Three steps to put your whole organization on AI
Create organizations and projects
Set up workspaces and projects by department or application, invite members, and assign roles.
Authorize models and allocate quota
Enable the allowed models for each project and allocate budget down the organizational hierarchy.
One key to start calling
Use a project key to call every major model format through a single gateway, with usage metered in real time.
Every adoption project stands on this infrastructure
ATP Petrichor isn't another tool to learn — it's the governance foundation built into every Horizon AI adoption project. From day one of the PoC, cost, permissions, and logs are already in place; going live just turns up the volume.
Explore our AI adoption services