What the built-in covers

claude --continue picks the most recent session in the current directory back up, and claude --resume lets you choose from a list. For "I closed the terminal an hour ago and want to keep going", that is the right tool and nothing here improves on it.

Where it stops

One directory

The list is scoped to where you are. Work done in a sibling repo, or in the same repo from a different path, is not offered.

One tool

It knows nothing about the session you ran in Codex or OpenCode, because that transcript is in a different format in a different directory.

Resume, not search

You pick from recent sessions. There is no way to ask which session touched a given file, or where a particular error was solved, or what was decided about a design six weeks ago.

You have to ask

Resuming is something you do deliberately. The assistant cannot check its own history mid-task, because it has no tool for that.

What replaces which

Nothing replaces the built-in. Keep using it for the same-day, same-directory case. The gap it leaves is recall, and recall is a search problem:

the part --continue has no answer for

$ chat-recall search "why did we drop the pooler"
  every session, every tool, ranked

$ chat-recall recent
  what was I working on, across all of them

And with the MCP server registered, the assistant runs those queries itself rather than waiting to be told.

The honest comparison

Same session, same day, same directory

Use claude --continue. It is free and it is already there.

Something from last month

Search. Resuming a list of recent sessions will not find it.

Something you did in another tool

Search. The built-in cannot see it at all.

Letting the assistant check for itself

MCP. That is the whole difference between a history you browse and a memory the agent uses.

Start free, no card How it works

You already have claude --continue. Here is what it does not do.