---
title: "AgentDrive vs Built-in resume"
description: "Claude Code --resume and codex resume continue a session in the same agent. AgentDrive syncs sessions so another agent or machine can pick up the work."
canonical: "https://agentdrive.sh/compare/built-in-resume"
---
# AgentDrive vs Built-in resume

> For the complete documentation index, see [llms.txt](https://agentdrive.sh/llms.txt). Every docs page is also served as markdown at its URL plus `.md`, and the full docs are in one file at [llms-full.txt](https://agentdrive.sh/llms-full.txt).

> Claude Code and Codex each resume their own sessions. AgentDrive syncs sessions so a different agent, or the same agent on another machine, can pick up the work.

Source: https://agentdrive.sh/compare/built-in-resume. Last checked 2026-11-18. All comparisons: https://agentdrive.sh/compare.md.

## When to pick which

Pick Built-in resume when: You stay in one agent on one machine and want the full conversation back exactly as it was.

Pick AgentDrive when: You switch agents mid-task, for example when you hit a usage limit, or move to another machine, and want the next agent to know where the last one stopped.

Use built-in resume to continue in the same agent, and AgentDrive when the next step happens in a different agent or on another machine. AgentDrive's hooks run alongside the agent and leave its own sessions and resume as they are.

| Feature | AgentDrive | Built-in resume |
| --- | --- | --- |
| Unit of storage | Sessions and files, versioned, with provenance | The agent's own saved conversation |
| Where it runs | Hosted workspace plus a local CLI and MCP server | Built into Claude Code (claude --continue, claude --resume) and Codex (codex resume, /resume) |
| Coding-agent sessions | Synced by hooks in Claude Code, Codex, OpenCode, Cursor and Antigravity; resumable from any of them | Resumed in the agent that recorded them; Codex's /import brings in recent Claude Code and Cursor chats |
| Storage location | AgentDrive R2 by default; your own R2 or S3 bucket on Team | Claude Code: JSONL transcripts under ~/.claude/projects, kept 30 days by default |
| Availability | Free plan; Pro $20/mo, Team $30/seat/mo | Included with each agent; the Codex CLI is open source (Apache-2.0) |

## About Built-in resume

- Claude Code: `claude --continue` (`-c`) loads the most recent conversation in the current directory, and `claude --resume` (`-r`) resumes a session by ID or name, or opens an interactive picker.
- Claude Code: `--fork-session` creates a new session ID when resuming, `--from-pr` opens the picker filtered to sessions linked to a pull request, and `--teleport` resumes a Claude Code on the web session in your local terminal.
- Claude Code stores each session locally as JSONL at `~/.claude/projects/<project>/<session-id>.jsonl`, with 30-day retention by default (`cleanupPeriodDays`). A resumed session restores the messages, tool calls and results.
- Inside Claude Code, `/resume` switches to an earlier conversation, and `/remote-control` lets you continue the local session from another device.
- Claude Code's `/import` brings configuration from Codex or Cursor on your machine: instruction files, MCP servers, commands, subagents and skills. Sessions are not on that list.
- Codex: `codex resume` continues a previous interactive session by ID or name, `--last` skips the picker and resumes the most recent chat from the current directory, and `--all` includes sessions from other directories. `/resume` does the same from inside a session, and `codex fork` copies a session into a new chat.
- Codex's `/import` brings in Claude Code or Cursor setup, projects and chats. The CLI looks for up to 50 chats from the last 30 days on the machine it runs on, and the desktop app can keep imported work in sync with the original agent.
- The Codex CLI is open source under the Apache-2.0 license.
- Claude Code's license reads: "All rights reserved. Use is subject to Anthropic's Commercial Terms of Service."

## Has either agent added cross-agent import?

Codex has. Its `/import` command copies recent Claude Code and Cursor chats into Codex, so you can open one there and continue. The CLI takes up to 50 chats from the last 30 days, and the desktop app can keep imported work updated from the original agent. Either way the work moves one way, into Codex, on the machine Codex runs on.

Claude Code's `/import` brings configuration from Codex or Cursor, such as instruction files, MCP servers and skills, not their sessions.

## What the next agent gets

Built-in resume restores the conversation itself, so the agent continues where it stopped. In Claude Code, a compacted session reopens from the compaction summary plus the messages after it.

AgentDrive gives the receiving agent a handoff brief instead: the goal taken from the first user message, the files touched, and as many recent turns as fit in 12,000 characters by default. The agent is told to check the files on disk before relying on it. In local mode, AgentDrive builds that brief from the same transcripts the agents keep on disk.

## Where Built-in resume works well

- No setup: resume ships with each agent and restores the full conversation, tool calls and results.
- Branching: `--fork-session` in Claude Code and `codex fork` let you try another approach without losing the original.
- Staying in one agent: when the next step happens in the same tool, nothing beats its own resume.

## What AgentDrive does differently

- AgentDrive resumes in every direction: a session synced from Claude Code can be picked up in Codex, OpenCode, Cursor or Antigravity, and the other way round. Codex's /import only brings work into Codex.
- Sessions sync to your workspace, so you can resume one on another machine signed in to the same account.
- Resume from inside the agent with `drive_resume`, a slash command, or `adrive resume`. MCP resume prefers a recent session in this project from a different agent.
- Each synced session sits in a versioned workspace with provenance: which agent wrote what, in which session.

## Frequently asked questions

### What is the difference between Built-in resume and AgentDrive?

Claude Code and Codex each resume their own sessions. AgentDrive syncs sessions so a different agent, or the same agent on another machine, can pick up the work.

### When should I use Built-in resume?

You stay in one agent on one machine and want the full conversation back exactly as it was.

### When should I use AgentDrive instead?

You switch agents mid-task, for example when you hit a usage limit, or move to another machine, and want the next agent to know where the last one stopped.

### Can I use AgentDrive alongside Built-in resume?

Yes. Use built-in resume to continue in the same agent, and AgentDrive when the next step happens in a different agent or on another machine. AgentDrive's hooks run alongside the agent and leave its own sessions and resume as they are.

## Sources

- https://code.claude.com/docs/en/cli-reference
- https://code.claude.com/docs/en/sessions
- https://code.claude.com/docs/en/commands
- https://learn.chatgpt.com/docs/developer-commands?surface=cli
- https://learn.chatgpt.com/docs/import
- https://github.com/openai/codex
- https://github.com/anthropics/claude-code/blob/main/LICENSE.md

