Your brand. Your customers. Your margin.

Ship predictive features inside your own product without standing up a data-science team. Phaeth Cortex is the forecasting layer you embed in your AI agent and ship under your own name: demand, revenue, churn and outcome prediction, called over a stateless API & SDK. We own the modeling; you ship the feature.


Embed it. Brand it. Bill it.

Cortex is infrastructure, not a logo on your customer's screen. You plug predictive intelligence into your agent, ship it as your own feature, and keep the relationship and the margin.

🧩
EMBED, DON'T REBUILD

Predictive intelligence as an API

Call Cortex over a stateless API & SDK (TS/Python), MCP-ready, from inside your own agent. We own the modeling, feature engineering, training, calibration, and you ship the feature your roadmap has been promising.

🏷️
WHITE-LABEL

Your brand, your customers, your margin

Your customers see predictions in your UI under your name, they never know Phaeth is behind the curtain. Usage-based pricing that grows with you, so the cost stays aligned with the value you capture.

🔧
WE MAINTAIN IT

No ML team required

Automatic feature-engineering, validation, and retraining on drift run inside Cortex. The model stays calibrated as the data shifts, without you hiring a permanent cost centre to keep it alive.


Two deployment modes. One API surface.

Some of your customers are fine in the cloud. Others, regulated, sensitive, air-gapped, won't let raw data leave. Pick the mode per deployment; only the base URL changes, your integration stays identical.

Enterprise SaaS

  • Hosted at cortex.phaeth.com
  • Multi-tenant, fully managed
  • Fastest integration: point at the base URL and go
  • We handle scaling, uptime, retraining
  • Best for SaaS, SMB, and consumer products

Local / VPC

  • On-prem or inside your own VPC
  • Single-tenant, isolated
  • Raw data never leaves: only encrypted structural metadata reaches the AI layer
  • Reproducible recipes: replayable byte-for-byte
  • Best for regulated industries and enterprise contracts

Both expose the same API surface.
Only the base URL changes.


Built local-first, from the ground up.

Whatever the deployment mode, the contract is the same: raw rows stay where your data lives, only encrypted structural metadata reaches the AI layer, and every model is a reproducible recipe you can replay.

LOCAL-FIRST METADATA ONLY SELF-HOSTABLE REPRODUCIBLE

Usage-based. No revenue share.

You pay per prediction, and the bill grows with your prediction volume: more customers and more API calls mean a proportionally larger Cortex bill, never a cut of your customer revenue. You set the price of the feature in your own tiers and keep the markup between what you charge and what you pay Phaeth. The cost stays a predictable line item that scales up and down with your business.


Ready to ship a predictive feature
under your own brand?

Get an API key and start integrating, with the full API surface and SDKs ready to go.

LOCAL-FIRST METADATA ONLY SELF-HOSTABLE REPRODUCIBLE