Background
- **Claude Code** is an agentic coding tool from Anthropic that lets you run Claude directly in your terminal to edit codebases, run commands, and manage software projects conversationally. Each "continue" starts a new turn and consumes tokens, which can add up fast.
- **Recall** is a plugin for Claude Code that records your session's context — past commands, file edits, decision logs — and automatically re-injects it into new continuations, so the AI doesn't forget what happened earlier and you waste fewer tokens re-explaining things.
- The problem it solves: when you resume a Claude Code session, the model often has no memory of prior steps, causing redundant work, repeated token spend on re-establishing context, and broken workflows.
- **Why it matters**: for developers using Claude Code intensively (especially on large projects), token costs and context loss are real friction points. Recall aims to reduce both by persisting and restoring session memory across restarts.