Transcript Lake

Documentation / CLI

transcript-lake show

Reconstruct one masked conversation in chronological order.

Invocation#

sh
transcript-lake [--data-dir <path>] show <session-id> [--include <types>] [--limit <n>] [--json]

Purpose#

Reconstruct one masked conversation in chronological order.

Inputs and options#

  • Required <session-id> must identify a session in the selected Lake.
  • --include <types> is a comma-separated list of user, assistant, thinking, tool_call, tool_result, meta, or hook_decision, or all; default is user,assistant.
  • --limit <n> keeps the oldest n events; without it every event of the session is shown. --json emits one report object.

Output and state effect#

Read-only. Human output includes session identity, span, turn counts, the selected types, full masked event text, and a rendered-versus-matched footer. JSON includes the same identity and event array; a --limit cut is always visible.

Exact refusals#

  • Wrong positional count: usage: transcript-lake show <session-id> [--include <types>] [--limit <n>] [--json].
  • Unknown session: unknown session "<session-id>": not present in the selected Lake (check the id or start the stream first).
  • Empty include list: --include needs at least one event type or "all".
  • Unknown type: unknown event type "<type>" (expected one of: user, assistant, thinking, tool_call, tool_result, meta, hook_decision, all).
  • Invalid limit: --limit must be a whole number of at least one; omit it to show every event of the session.

All long flags are exact and single-use. A value flag never consumes another --flag. Global --data-dir may appear anywhere in the invocation, must appear once, and selects an absolute root.

Back to the complete CLI command tree.