Skip to main content

​
Getting started

​
Run it

That is the whole install. Turnlog starts a server on 127.0.0.1, prints a URL, and opens your browser:
The Codex and Cursor lines appear only if those histories exist. There is nothing to configure either way — Turnlog finds what is there.
The port is random on every run and the ?token= is generated fresh each launch, so copy the whole UI: line if the browser does not open by itself. Both are deliberate: see Privacy.
Prefer a global install? npm i -g turnlog, then turnlog. Node 22.13 or newer.

​
The first run

The first index is a one-time pass over every session file you have. It streams each file rather than loading it — sessions run to hundreds of megabytes — and sessions appear in the UI as they are parsed, so you can start reading before it finishes. After that, indexing is incremental: Turnlog remembers each file’s byte offset and reads only what is new. A file watcher keeps the UI live while an agent is still writing, so a session you are running right now updates within about a second.

​
Where things live

Override the index location with TURNLOG_DATA_DIR. Deleting the index costs you nothing but a re-index — your annotations (pins, names, notes, tags, bookmarks, saved searches) survive a rebuild, because they live in tables the rebuild does not touch.

​
Try it without your own history

Runs the real UI against bundled sample sessions in a scratch index. Your own history is not read and not touched, and the banner says so for as long as it is running.

​
Next