Our mission

The context of your AI agents
belongs to you.

Tramya gives your AI coding agents a durable memory. We start from a simple conviction: what your assistants learn about your work — decisions, history, conventions — should stay with you, and be able to move from one agent to the next without getting lost.

Why local-first

The memory of your work doesn't belong to a cloud.

Your data, your machine

Coding agents accumulate valuable, intimate context. Leaving it on third-party servers means losing control of it. Tramya indexes that context where it already lives: on your disk.

Continuity across agents

You switch tools — Claude Code, Codex, Cursor, Gemini — but your project doesn't. Tramya hands the real context from one agent to the next over MCP, without starting from scratch.

The Cloud, only if you want it

Sync exists, but it's optional and end-to-end encrypted. We design the Cloud as an encrypted bridge between your machines, never a warehouse that reads your content.

Restraint by default

No product telemetry, no silent collection. A memory tool has to earn trust: ours does as little as possible with your data.

What we build

A memory, not yet another assistant.

The founder

Arthur Ferreira

Arthur Ferreira — founder of Tramya.

“I built Tramya because I spent my days repeating the same context to different agents — Claude Code in the morning, Codex in the afternoon, another one the next day — and every time it started from scratch: the decisions made, the files touched, the dead ends already explored.

Tramya keeps that thread, on your machine, and hands it from one agent to the next. Local-first on principle: your code context is the most intimate thing you have — there's no reason for it to leave for a server.”

A question, an idea, a bug? Write to me directly: support@tramya.com.