Explorer

Explorer is an interactive terminal UI for building RecordStream pipelines. Instead of writing shell one-liners and re-running them every time you tweak something, you build your pipeline step by step — adding stages, inspecting output, and iterating — all inside a single terminal session.

Why Explorer?

The classic recs workflow is powerful:

recs fromcsv --header data.csv \
  | recs grep '{{status}} >= 500' \
  | recs collate --key host -a count \
  | recs sort --key count=-n \
  | recs totable

But exploratory data work means constant iteration — change the filter, re-run, check the output, add a sort, re-run again. With shell pipelines, every tweak means hitting up-arrow and editing a long command.

Explorer replaces that cycle with a live, interactive loop:

1. Add a stage
2. See its output instantly
3. Tweak, reorder, or delete stages
4. Export the finished pipeline as a shell script

The Interface

Explorer uses a split-pane layout: a pipeline list on the left, a data inspector on the right.

+==============================================================================+
| recs explorer                 input: access.log (5000 rec)   fork: main  [?] |
+=========================+====================================================+
| Pipeline                | Inspector: grep (2847 records, cached 3s ago)      |
|                         |                                                    |
|   1 fromre       5000   | #   ip             status  time    path      size  |
|     '^(\S+)...'         | 1   192.168.1.1    200     161..   /index    1234  |
|   2 grep     +   2847   | 2   10.0.0.5       200     161..   /api/u    5678  |
|  >  status=200     <--  | 3   192.168.1.1    200     161..   /style    910   |
|   3 sort     !   ----   | 4   172.16.0.12    200     161..   /favic    234   |
|     --key time=n        | 5   10.0.0.5       200     161..   /api/d    8901  |
|   4 collate  !   ----   | 6   192.168.1.100  200     161..   /image    4567  |
|     --key host          | 7   10.0.0.22      200     161..   /js/ap    2345  |
|     -a count            | 8   172.16.0.12    200     161..   /api/c    6789  |
|   5 totable  !   ----   | 9   192.168.1.1    200     161..   /fonts    123   |
|     (output)            | 10  10.0.0.5       200     161..   /login    4560  |
|                         | ...                              (2847 total)      |
+-------------------------+----------------------------------------------------+
| a:add d:del e:edit u:undo x:export  f:fork v:vim ?:help q:quit | undo:3     |
+==============================================================================+

Left panel: Your pipeline stages, with cursor navigation. Each stage shows its operation name, arguments, cache status, and record count.

Right panel: The output of the currently selected stage. Move the cursor and the inspector updates immediately — cached results appear instantly, uncached stages execute on the fly.

Status bar: Quick-reference keybindings and undo count.

Shell Pipelines vs Explorer

	Shell Pipelines	Explorer
Iteration	Edit command, re-run, check output	Move cursor, see output instantly
Visibility	See one stage's output at a time	Jump between any stage to inspect
Experimentation	Copy-paste to try alternatives	Fork your pipeline into branches
Undo	Hope you remember what you had before	u to undo, Ctrl+R to redo
Export	Already a shell command	x to copy as shell script
Sessions	Terminal history	Named sessions, auto-saved to disk

Explorer isn't a replacement for shell pipelines — it's where you figure out what your pipeline should be. Once you're happy, export it and drop it into a script.

Next Steps

• Getting Started — Open your first file and build a pipeline
• Keyboard Reference — Every key, organized by context
• Feature Guide — Deep dives into forks, sessions, quick actions, and more