omni stats
Token savings analytics, read from your own database.
omni stats
Leads with session lifetime, how many commands a session carries before the host closes it. The distillation percentage below it is a diagnostic for one host’s pipeline, not a product claim.
Flags
| flag | effect |
|---|---|
--detail | Full breakdown: commands, routes, sessions, agents |
--hour, -H | Scope to the last 60 minutes |
--day, --today, -d | Today only |
--week, -w | Last 7 days |
--month, -m | Last 30 days, the default |
--all-commands | Every command, not just the top ones |
--project | Break down per project path |
--context | Context composition signals |
--rerun | Which distillers cost a re-run |
--share | A copy-pasteable summary of your measured savings |
--card | Write that summary as an image, sized for social posts |
--json | Machine readable |
--help, -h | Help |
--rerun is the one to know
Reduction percentage cannot tell you whether a distiller removed something the agent then had to fetch again. If it did, the reduction was a deferral, not a saving. This flag is the check that percentage cannot make.
Traps
Terminal rows are not tokens. Output written to a TTY is read by a human, not a
model. On one installation those rows were 73% of every byte OMNI claimed to have
saved. stats excludes them now, and so does the benchmark harness, but anyone
querying ~/.omni/omni.db directly has to filter by agent_id themselves.
A high number deserves suspicion. The worst defects in this project reported the
highest reductions, because deleting the answer compresses very well. Pair any figure
with omni diff on a real command.
A low number is usually right. Around 97% of calls save nothing because there was nothing to save.
--share and --card cannot drift from the report. Both read the same
aggregation as omni stats itself, which was a deliberate choice after an earlier
version computed them separately.