elephant

Codebase recovery from memory reconstructs files and directories shown or created earlier in your conversation. Elephant restores your full codebase or selective files by extracting them from Claude's conversation history, Read/Write/Edit tool logs, Bash output, and generated code, then writes them back to your specified recovery path.

Yes, selective restoration lets you recover specific files or directories from memory rather than the full codebase. Specify the exact paths or file names you need recovered, and Elephant extracts only those items from conversation history and tool outputs.

Memory-based recovery applies when files are accidentally deleted, destructive commands wipe critical code, Git history is unavailable or corrupted, or you need to checkpoint your current state before risky operations. It preserves work that exists only in conversation context.

Memory checkpointing captures your current codebase state as a snapshot before executing potentially destructive commands. This snapshot stays in conversation memory and can be fully or selectively restored later if something goes wrong.

Yes, memory-based recovery works independently of Git availability. If your Git history is unavailable or corrupted, Elephant reconstructs files from Claude's conversation memory, tool execution logs, and code generated or displayed during your session.

Elephant restores code from multiple sources: Read/Write/Edit tool histories, Bash command output, user-provided code, generated code within the conversation, and memory checkpoints. It synthesizes these sources to reconstruct the full or partial codebase state.