Hey everyone 👋 I kept hitting the same wall. I'd build up a rich session with a coding agent, full context, reasoning, tool calls, then hit a usage limit, switch machines, or want a second model's opinion, and have to start over and re-explain everything. Every agent locks your session in its own on-disk format, so there was just no way to move it. So I built AgentSync. It reads each agent's native session format, converts it through a single canonical model (built on the open-source txcript engine), and writes it back out for another agent. That means a Kiro to Claude to Codex handoff keeps your whole conversation, reasoning, and tool history. Add the cloud layer with push/pull and the same session follows you across machines. It's a single Node CLI, local-first (nothing leaves your disk until you push), and you install it in one line: npm install -g @juli_p/agentsync. Docs are at https://agentsync.trailmark.onli... Would love your feedback, especially on which agents you'd want supported next. 🙏

AgentSync - Context Transfer in seconds

Move any coding-agent session anywhere

Move any coding-agent session anywhere