APLEXICA

A story about getting locked in and deciding to do something about it.

The story

A working software engineer spent months teaching one AI agent how their codebase worked, how the team made decisions, how they liked to write. Then a better agent shipped. Two weeks of evaluation later, the original agent still knew everything; the new one knew nothing. They went back to the original.

That same person started watching the agent space. Every few weeks something new launched. Every time, the same calculus: do I want to start over? The answer was always no. They were locked in — not by a contract or by a missing feature, but by the cost of moving their own context.

That cost is a tax on every developer running AI agents. Aplexica removes it. Your memories, your skills, your conversation history — all yours, portable, and readable by any agent you choose.

Your agent state is yours — by authorship, not by permission.
Founding principle 2026-04 · /docs/principles.md

What we believe

Aplexica's job is to make switching agents, running several in parallel, and trying new ones a non-event. The Open Source edition stays free, forever. The Cloud business funds the open-source work. The keys, the architecture, and the data model stay user-controlled.

Want to follow along? The Open Source edition is shipping now.

Subscribe to the blog or watch the GitHub repo.