Manifesto

Why It Matters

LLMs are powerful, but they’re stateless. They forget everything unless you keep re-prompting them with the same context over and over. This makes it impossible to build apps that need continuity: apps that must remember user preferences, track ongoing work, update states across tools, or understand what happened last week without you manually re-explaining it.

I ran into this firsthand while building a unified inbox for founders (Gmail + Slack + LinkedIn + and more). I needed the system to automatically remember what mattered to each founder, update context over time, and understand ongoing threads. But current LLMs couldn’t hold or update that memory. The product fundamentally couldn’t exist without solving the memory problem.

Memory becomes essential when an app needs:

• State that updates itself
  (e.g., sales progress in Salesforce, project movement in Linear, hiring pipelines in Notion)

• Continuity across events
  (voice logs, email threads, Slack discussions, meeting notes)

• Cross-tool understanding
  (Gmail ↔ Slack ↔ GitHub ↔ CRM workflows)

• Accurate company knowledge
  Startup “organizational memory” only works if the system can actually remember evolving facts, people, and decisions.

Without reliable, persistent memory, all “AI workflows” break the moment the user stops talking.

By giving apps Gmail/Slack/Linear/Salesforce-level long-term memory, you unlock entire product categories that simply aren’t possible today:
tools that automatically update themselves, understand what’s happening across channels, and maintain a consistent, evolving picture of the user or the company.

Momo is starting with Gmail memory first, because once you can persist and update what happens in email, you’ve already captured the single most important source of truth for most people and companies. From there, the memory layer extends naturally to the rest of the stack.