What format is ZCode history stored in?

Unlike Claude Code, which writes one readable .jsonl text file per session, ZCode stores its history in SQLite databases under ~/.zcode — for example a session database under ~/.zcode/cli and a task index alongside it. You can inspect them with any SQLite tool, but they are an internal storage format, not a documented export.

Can I back up my ZCode history?

Yes. Because everything lives under ~/.zcode, copying that folder (with ZCode closed, so no database is mid-write) backs up your sessions and task history, and copying it back restores them. Treat the copy as sensitive: it can contain code, paths, and details from your projects.

Can I grep ZCode sessions like Claude Code sessions?

Not directly. Claude Code sessions are plain-text .jsonl files, so a text search across ~/.claude/projects works. ZCode history is in SQLite databases, so tools like ripgrep cannot usefully search inside them — you need a SQLite-aware tool, or an indexer like LLMnesia that reads the history and makes the conversations searchable by keyword.

How do I search my ZCode session history?

ZCode's own interface shows your past sessions, but for keyword search across everything on disk you need to either query the SQLite databases yourself or index them with a tool. LLMnesia indexes your ZCode prompts and the assistant's replies locally — never subagent runs, commands, or tool output — and makes them searchable alongside your Claude Code, Codex, and web AI chats.

Where Does ZCode Store Session History?

ZCode stores your session history locally, under ~/.zcode in your home directory — inside SQLite databases, not readable text files. Like Claude Code, ZCode keeps your history on your machine rather than on a server. Unlike Claude Code, there is no folder of one-file-per-conversation transcripts you can open in a text editor, which changes what finding and searching an old session looks like. Here is how the storage is laid out.

LLMnesia⌘ ⇧ 9

mediterranean diet inflammation

30 indexed · ↑ for previous

Mediterranean diet effects on chronic inflammation markers

Multiple cohorts show a mediterranean diet lowers CRP and other inflammation markers.

Perplexity6 days ago

Build a mediterranean diet meal plan low in inflammation triggers

A mediterranean diet built on olive oil, fish, and greens limits inflammation triggers.

ChatGPT2 weeks ago

Summarise this paper on mediterranean diet and inflammation

The paper finds mediterranean diet adherence correlates with reduced systemic inflammation.

Claude3 weeks ago

The short answer

Your ZCode data lives here:

~/.zcode/

On macOS and Linux that expands to /Users/you/.zcode or /home/you/.zcode. On Windows it sits under your user profile folder. It is a hidden folder (the leading dot), so you may need to enable "show hidden files" in your file manager to see it.

Inside it, the pieces most relevant to your history are:

A session database — under ~/.zcode/cli/, holding your sessions and messages (tables for sessions, session entries, and messages sit alongside internal bookkeeping like tool usage and model stats).
A task index — ~/.zcode/v2/tasks-index.sqlite, indexing ZCode's task history so the app can list and reopen past work.
Configuration — ~/.zcode/cli/config.json and its neighbours, which is also where ZCode's MCP servers are configured.

The exact filenames can shift between ZCode versions, but the shape is stable: everything ZCode remembers about your work lives in this one folder, in databases rather than plain text.

How this differs from Claude Code

The comparison matters because the two tools make opposite trade-offs in storage. Claude Code writes one .jsonl file per session under ~/.claude/projects — plain text, one JSON object per line, grep-able with any text tool. ZCode keeps its history in SQLite databases under ~/.zcode, which the app reads and writes directly.

Practical consequences:

Text search tools do not work. ripgrep across ~/.zcode will not find a phrase you remember saying, because the words are inside database pages, not plain text.
No per-session file to open. You cannot double-click one file to read one conversation; you need a SQLite-aware tool.
One folder still backs it all up. Copy ~/.zcode (with ZCode closed) and you have your history; paste it back on a new machine to restore it.

Reading the history yourself

If you want to look inside, any SQLite tool works — the sqlite3 command line, or a GUI like DB Browser for SQLite. Two precautions:

Work on a copy. These are live databases with write-ahead logs; open a copy rather than the file ZCode is using.
Expect internals, not a transcript. The tables mix your prompts and the assistant's replies with internal records — tool usage, model stats, workflow state — so raw queries return noise alongside the conversation.

For a one-off "did I really say that" check, that is fine. For anything you do more than once, it gets old fast.

How this compares to other AI tools

Tool	Where history lives	Can you back it up yourself?
ZCode	`~/.zcode` on your machine (SQLite)	Yes — copy the folder
Claude Code	`~/.claude/projects` on your machine (`.jsonl` files)	Yes — copy the folder
Claude (web/desktop)	Anthropic's servers, shown in the sidebar	Only via an account data export
ChatGPT	OpenAI's servers	Only via an account data export
Gemini	Google's servers, tied to your account	Via Google Takeout

The local-first approach — shared by ZCode and Claude Code — means your history does not depend on an account staying active, and a folder copy is a complete backup. The trade-off is that nothing indexes it for you, and ZCode's database format makes casual searching harder than Claude Code's plain text files.

Making it searchable

This is the gap tools like LLMnesia fill. It indexes your ZCode sessions locally — only your prompts and the assistant's replies, never subagent runs, shell commands, or tool output — so results read like a clean transcript instead of raw database rows. The same search also covers your Claude Code and Codex sessions and your ChatGPT, Claude, and Gemini chats, and its free MCP connection lets ZCode itself search that history when you ask it a question about past work. Everything stays on your device.

If you are trying to track down one particular session right now, the step-by-step guide to finding an old ZCode session works through each method in order.