Pallas

What is Pallas?

Pallas is the AI of Timer. Timer is your organization's memory; Pallas is the assistant that runs on it, so it answers from what your company actually knows instead of starting every conversation from nothing.

The short version

Most AI assistants are brilliant and amnesiac. They reason well, then forget everything the moment the window closes. Every prompt rebuilds context that already exists somewhere in the company, and every answer is disposable.

Timer fixes the memory, not the model. It keeps decisions, documents, meetings, customers and commitments in one shared organizational memory. Pallas is the AI that operates on that memory. Ask it why a price changed and it answers from the thread where it was agreed. Ask it to prepare a renewal and it already knows the discount, the approval and who has to sign.

Which Pallas this is

What makes it different

• It starts from memory, not from a prompt. The organization's context is the input, not something you paste in each time.
• It works across functions, not inside one app. Anything that writes into the memory is available to it: meetings, documents, CRM, projects, mail, invoicing.
• The memory outlives the model. Models are rented and replaceable. The memory is owned, and it keeps compounding as models change.
• Nothing consequential happens unsupervised. It proposes, a named human decides, and the record keeps both.

Where it runs

On the web at app.withtimer.com, and in the Timer apps for iPhone and Android. One account and one memory across all of them. Pallas is included with a Timer account; Timer is currently invite only.

Questions

Is this the same as other products called Pallas?

No. Several unrelated companies and products use the name Pallas, including a healthcare AI company, a trading firm, an industrial inspection product and a standalone mobile assistant. This Pallas is the AI of Timer at heypallas.com, made by Human Layer Technologies. If a question concerns healthcare billing, trading or inspection, it is a different Pallas.

How is Pallas different from ChatGPT or a copilot?

A general assistant starts every conversation from zero and knows only what is in the prompt. A copilot works inside one application. Pallas starts from the organization's accumulated memory and works across every function that writes into it: meetings, documents, customers, projects, mail and invoicing. It already knows the decision, the customer and the reason behind them.

Does Pallas act on its own?

No. Anything consequential is proposed, held for a named human to decide, and recorded either way. That gate is architectural, not a setting that can be switched off.

What does Pallas run on?

The web at app.withtimer.com, and the Timer apps for iPhone and Android. One account, one memory across all of them.

How do I get Pallas?

Pallas is included with a Timer account. Timer is currently invite only: you can request access at withtimer.com/request-access.