Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add commands/erhangundogan/videre/statsgit clone --depth 1 https://github.com/erhangundogan/videreWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/erhangundogan/videre/stats)<a href="https://agentmods.dev/commands/erhangundogan/videre/stats"><img src="https://agentmods.dev/badge/commands/erhangundogan/videre/stats.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00012 | $0.02692 |
| Opus 5 | $0.00006 | $0.01346 |
| Sonnet 5 | $0.00002 | $0.00538 |
| Haiku 4.5 | $0.00001 | $0.00269 |
Grade A, and why
stats scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 241 lines — stays where its author put it; the contents beside it link to each section on GitHub.
videre stats # library totals and what has run recently
videre stats --json # print one JSON object instead
videre stats --check # exit non-zero if anything failed or crashed (for cron)
videre stats --db ~/photos.db # use a specific database
Reading the output
Library: 70601 file(s) (398.9 GB), 57142 photo(s), 13459 video(s)
Duplicates: 0 group(s), 0 file(s), 0 B wasted
Faces: 58555 detected, 86 people named
Marks: 312 rated, 40 picked, 12 labelled, 128 liked
Embeddings:
google/siglip-base-patch16-224 70588 768-dim 128.3 MB
google/siglip-so400m-patch14-384 70587 1152-dim 189.6 MB
google/siglip2-base-patch16-384 70588 768-dim 128.3 MB
By type:
mov video/quicktime 12564 165.4 GB
jpeg image/jpeg 43524 142.4 GB
mp4 video/mp4 643 62.0 GB
heic image/heic 10395 20.8 GB
png image/png 3211 6.5 GB
mov video/mp4 251 909.5 MB
mp4 video/quicktime 1 462.5 MB
dng image/tiff 9 433.9 MB
gif image/gif 2 4.8 MB
png image/jpeg 1 43.7 KB
Disk use:
embeddings 446.3 MB
database 161.8 MB
thumbnails 84.1 MB (rebuildable)
place names 11.4 MB (rebuildable)
database journal 32.0 KB (rebuildable)
total 703.7 MB (95.6 MB of it rebuildable)
Pipeline status:
scan success last_run=2026-08-09 16:17:43 duration=0ms
faces interrupted last_run=2026-08-03 10:11:45 duration=9m 6s
embed success last_run=2026-08-17 13:29:01 duration=21s
classify success last_run=2026-08-17 13:29:30 duration=1.5s
dedupe success last_run=2026-08-05 11:35:38 duration=132ms
fix-dates success last_run=2026-07-31 10:33:15 duration=50ms
prune success last_run=2026-07-31 21:02:36 duration=479ms
locations success last_run=2026-08-18 09:32:47 duration=1m 17s
Real output from a 70,000-file library, so the awkward rows are the interesting ones.
By type groups the library by file extension, with the mime type beside it
and the largest first. These are the values --ext and --mime take, so this is
where to look before scoping a run.
Extension and mime disagree more often than people expect, and both are shown
rather than reconciled. Three rows above are the same file types filed under a
different name: 251 .mov files holding MP4 video, one .mp4 holding QuickTime,
and one .png that is really a JPEG. Reconciling them would mean choosing which
of the two to lie about, so --ext mov and --mime video/quicktime select
genuinely different sets.
Disk use is what videre itself stores, not your photos, largest first.
Entries marked (rebuildable) cost only time to recreate: thumbnails are
re-converted on demand, place names ship with videre, and the database journal is
transient. The database and embeddings are not rebuildable - an embedding run
takes hours, and the database cannot be recreated without rescanning.
Note the proportions above: 703.7 MB of videre data describing 398.9 GB of photos, and only 95.6 MB of it disposable. Embeddings dominate because that library has three models prepared.
:::note[Embeddings are counted per library]
They live in a separate directory per database, so stats reports only the ones
belonging to the database you asked about. If other libraries share the same
VIDERE_HOME, theirs appear as a separate
embeddings (other libraries) row - real disk use, but not this library's.
:::
Duplicates counts exact, byte-identical copies only. "Wasted" is what you
would reclaim by deleting all but one of each group, which is what
videre dedupe proposes.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 3d ago First seen · 241 lines · 12 tokens per session scan A c5fdeb67ca62
stats is a command published in the GitHub repository erhangundogan/videre (2 stars, last pushed 5d ago), licensed Apache-2.0. It adds 12 tokens to every session and 2,692 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other commands, from other repositories
deslop
Check the diff against main, and remove all AI generated slop introduced in this branch.
contrib
Profile how top GitHub contributors work — ingest, build personas, and synthesise a capability model + learnable playbook.
memory-init
Initialize local memory store + vector index for this project.
graph
Generate an interactive visual graph of your memories. Powered by graphify (github.com/safishamsi/graphify, MIT).
git
Git operations with intelligent commit messages and workflow optimization.
checklist
Generate a custom checklist for the current feature based on user requirements.