Jeeves delivers a terminal interface for listing, searching and resuming AI agent conversations. Primary source states it supports Claude Code and Codex, allows content and regex search, split-pane previews and direct resume into the agent. (https://github.com/robinovitch61/jeeves)

GitHub documentation cites default 30-day Claude cleanup and instructs users to set "cleanupPeriodDays": 99999 in ~/.claude/settings.json for persistence; tool is built with Charm libraries and robinovitch61 viewport bubble. (https://github.com/robinovitch61/jeeves) Related Anthropic computer-use announcement notes long-running agent sessions require reliable history access. (https://www.anthropic.com/news/3-5-models-and-computer-use) OpenAI o1 system card similarly identifies context retention as scaling bottleneck. (https://openai.com/index/openai-o1-system-card/)

Primary sources show Jeeves addresses patterns seen in LangChain persistence modules and Aider's session commands; original README omits explicit linkage to these developer workflows now common in production agent use.