How to resume a Claude Code session (even on another machine)
Resume a Claude Code session with --continue, --resume, or /resume, find sessions from other folders, and pick up work on another computer with teleport.
You close the laptop mid-refactor. Next morning you're on the desktop at home, open Claude Code in the same repository, and it has no idea what you were doing. The session exists; it's just on the other machine.
This guide covers resuming on the same machine, finding sessions that don't show up, and moving work across computers, from Anthropic's sessions documentation, checked on 16 November 2026.
Resume on the same machine
| Command | What it does |
|---|---|
claude --continue | Reopens the most recent session in the current directory |
claude --resume | Opens the session picker |
claude --resume <name> | Resumes a named session |
claude --resume <session-id> | Resumes that session, from any directory |
claude --resume <path>.jsonl | Resumes the transcript at that absolute path |
claude --from-pr <number> | Opens the picker filtered to sessions linked to that pull request |
/resume | Switches session from inside Claude Code |
Name sessions as you go with /rename auth-refactor, and claude --resume auth-refactor finds them later without the picker.
When a session doesn't show up
Claude Code stores sessions per project directory, and the picker starts narrow. It shows sessions from the current worktree, plus sessions elsewhere that added this directory with /add-dir.
- Started in another worktree or folder: press
Ctrl+Win the picker for all worktrees of the repository, orCtrl+Afor every project on the machine. - Started with
claude -p, the Agent SDK, or a/loopfirst prompt: these are left out of the picker and--continue, butclaude --resume <session-id>still works. - Older than 30 days: transcripts are cleaned up after 30 days by default. Raise
cleanupPeriodDaysinsettings.jsonif you need longer.
claude --resume <session-id> searches the current project first, then every other project on the machine, so it finds a session wherever it started.
Resume on another machine
Sessions live on the machine that ran them, at ~/.claude/projects/<project>/<session-id>.jsonl. --resume reads only that local history. You have three ways to cross machines.
1. Start in the cloud, teleport anywhere
A cloud session runs on Anthropic's infrastructure, not your laptop. Start one with claude --cloud "task description" or at claude.ai/code. From any machine with a checkout of the same repository:
claude --teleport
That opens a picker of your cloud sessions; claude --teleport <session-id> goes straight to one. Teleport checks you're in the right repository, fetches and checks out the session's branch, and loads the conversation. Your working directory must be clean, and the branch must be pushed.
The handoff is one-way: you can pull a cloud session into a terminal, but you can't push an existing terminal session to the cloud. Teleport needs a claude.ai account sign-in, not an API key.
2. Steer the running session with Remote Control
If the first machine stays on, run /remote-control in the session, or start with claude --remote-control. You can then continue it from claude.ai/code in a browser on another computer, or from the Claude app on your phone. Claude keeps running on the original machine, with its files and MCP servers. Remote Control needs a Pro, Max, Ultra, Team, or Enterprise plan, and on Team and Enterprise an Owner must turn it on first.
3. Copy the transcript
Copy the .jsonl file to the second machine, check out the same repository and branch, and resume it by path:
claude --resume /path/to/<session-id>.jsonl
This brings the conversation, not the code. Push and pull your branch separately. Anthropic notes the file format is internal and changes between versions, so keep both machines on the same Claude Code version.
What a resumed session keeps
A resumed session restores the conversation history, the model, and an active goal. Resuming from a terminal with --continue or by ID or name usually restores the permission mode too, with exceptions: a session that ended in bypassPermissions doesn't resume in it.
It doesn't restore launch flags. Pass these again: --mcp-config, --plugin-dir, --add-dir, --settings, and --fallback-model. Background Bash commands and subagents that were running don't restart; they show up as a note that they didn't finish.
Resume or start fresh
Resuming a long conversation after a long break can be expensive, because the prompt cache has expired and the whole history is sent again. On Pro, Max, and Ultra plans, Claude Code may show a "Resume this conversation?" dialog with an estimate. If the next task doesn't need the old conversation, start fresh and keep the important parts in CLAUDE.md or auto memory instead. See how to reduce Claude Code token usage.
What a session can't carry
A session holds one person's conversation. It doesn't reach a teammate's machine, and it fades as conversations get compacted or cleaned up. Decisions made along the way, such as "we're keeping the v1 endpoint until March", need a home that outlasts the session.
Where Lockstep fits
Lockstep keeps those decisions outside any single session. It captures them from PRDs, Slack, and pull requests, routes each to its owner for approval, and briefs every Claude Code session at the start, on any machine and for any teammate, with the approved decisions that apply to the work.
See how it works.
Sources checked 16 November 2026
- https://code.claude.com/docs/en/sessions
- https://code.claude.com/docs/en/claude-code-on-the-web
- https://code.claude.com/docs/en/remote-control
Related reading
For what Claude remembers across sessions, see how Claude Code memory works. For parallel sessions on one machine, see Claude Code worktrees. For session commands, see the Claude Code slash commands cheat sheet. For moving sessions between the terminal, desktop app, and VS Code, see Claude Code desktop vs CLI vs VS Code. For cloud sessions in full, see Claude Code cloud sessions. For Remote Control in full, see Claude Code Remote Control.
Frequently asked questions
- How do I resume a Claude Code session?
- Run claude --continue to reopen the most recent session in the current directory, or claude --resume to choose from a picker. Inside a running session, /resume switches to another conversation. You can also pass a session name or ID, as in claude --resume auth-refactor.
- Can I resume a Claude Code session on another computer?
- Not directly with --resume, which reads this machine's local history. Start the work as a cloud session and pull it into any terminal with claude --teleport, use Remote Control to steer the session still running on the first machine, or copy the .jsonl transcript and run claude --resume with its absolute path.
- Why doesn't my session show up in claude --resume?
- The picker shows sessions for the current worktree by default. Press Ctrl+W to widen to all worktrees of the repository or Ctrl+A for every project on the machine. Sessions started with claude -p, the Agent SDK, or a /loop first prompt are left out, but resume by session ID still works.
- Where are Claude Code sessions stored?
- As JSONL files at ~/.claude/projects/<project>/<session-id>.jsonl, where <project> is the working directory path with non-alphanumeric characters replaced by dashes. They're kept for 30 days by default; change that with cleanupPeriodDays in settings.json.
- Does a resumed session keep its MCP servers and permission mode?
- It restores the conversation, model, and usually the permission mode. Launch flags such as --mcp-config, --plugin-dir, --add-dir, and --settings are not restored, so pass them again when you resume.
Keep reading