Troubleshooting

Start with --verbose

cca --verbose

  Claude Code usage · all time · 13 sessions across 3 projects

  Tokens   input  output  cache 5m  cache 1h  cache read
          144.5K  673.5K    503.0K      6.6M       52.0M

  Model                      tokens    cost
  claude-opus-5               20.3M  $37.87
  claude-sonnet-5             20.3M  $15.97
  claude-haiku-4-5-20251001   19.4M   $6.99

  Cost    $60.82     at API list prices
  Water   ≈ 18.0 L   about 2 minutes of shower

  ─ Cost is what this usage would cost at API rates, not what you were billed;
    Claude Code on a subscription draws from your plan allowance instead.
  ─ Water assumes 0.30 mL / 1k tokens, a rough estimate. See README.

  Scan · 18 files, 450 lines
                            count
  usage records kept          340
  duplicates dropped  110 (24.4%)

This block makes the headline numbers auditable: what was scanned, what was kept, and what was dropped and why.

The duplicates figure is the one to look at first. The same assistant message is replayed into several files by resumed sessions and compaction, so a ratio around half is normal and healthy. A ratio of zero would suggest deduplication is not working.

Common messages

claude directory not found

cca: claude directory not found: /home/you/.claude/projects
If Claude Code stores its data elsewhere, point cca at it with --claude-dir

Nothing exists at that path. Either Claude Code has never run, or its data is elsewhere:

cca --claude-dir /path/to/.claude

exists but holds no transcripts yet

The directory is there but empty. Normal for a fresh install — run Claude Code once and try again.

No usage found

The directory parsed fine and contained no usage-bearing records. If you passed a time range, try widening it; --verbose reports how many records the window excluded.

N models had no rate

─ 1 model had no rate and was excluded from the cost total: claude-future-9.
  Add it to pricing.json to price it; see README.

cca will never guess a rate. The tokens still count; only the dollars are missing. Add the model to a pricing override.

--json and --csv are mutually exclusive

Each describes one output shape. Pick one.

this command already sets its own time range

cca today --since 7d asks for two different windows. Use cca --since 7d.

Numbers look too high

Almost always deduplication. Check --verbose: if duplicates dropped is zero on a corpus with resumed sessions, something is wrong — please open an issue.

Numbers look too low

Three common causes:

1. A time range is active. Check the header line.
2. --no-sidechains is set, in a flag or in your config file. Check cca config.
3. The transcript is incomplete. Claude Code prunes history, and does not guarantee that every billed turn was recorded. See Accuracy and limits.

Totals changed between runs

Expected. Claude Code prunes and rotates transcripts, so historical figures can fall over time. cca reports what is on disk now; it is not an append-only ledger.

Output looks garbled

Your terminal is mangling the box-drawing characters:

cca --ascii

For escape sequences appearing literally, disable styling:

cca --no-color
# or
NO_COLOR=1 cca

Colour is missing when piping

By design — cca disables styling when output is not a terminal. To keep it while paging:

CLICOLOR_FORCE=1 cca | less -R