Provenance

Every number we publish, and where it comes from

There are 20 of them. Each carries a label saying what kind of number it is, and most carry the artifact they come from and the date that artifact last changed. No figure on this page is typed by hand — the prose around them is.

The numbers on a data company's website are the part a buyer can least easily check, so each figure here names what it is computed from.

The four labels

Every figure we publish is one of these four, and the label is part of the figure rather than a comment on it.

`measured`	computed from a committed artifact named in `source`
`derived`	computed from measured values
`target`	intended, not achieved — never publishable as inventory or capability
`none`	a real, deliberate zero

target is the one to watch. A target is what a capture brief is scoped to produce. It is not footage we hold, and it may never be read as availability.

Delivery formats: what we do and do not offer

Four containers and our own JSON schema ship. MCAP and WebDataset do not. What arrives, and in what container.

12 figures

Measured

Computed from a committed artifact, named in the source column. Re-run the command below and you get these numbers back.

Figure	How it is arrived at	Source artifact	Source last changed
1 end-to-end annotation runs `validated_runs`	count of *_Annotation.json files tracked by git in the engine's sampledata/	`fourthuman_engine_api/sampledata/Egocentric_Sample_2_Annotation.json`	2026-09-11
1,843 frames `reference_run_frames`	tracks.hand_control.total_frames_annotated	`fourthuman_engine_api/sampledata/Egocentric_Sample_2_Annotation.json`	2026-09-11
30 fps `reference_run_fps`	fps field	`fourthuman_engine_api/sampledata/Egocentric_Sample_2_Annotation.json`	2026-09-11
294 entries `reference_run_kinematic_entries`	len(tracks.kinematics_3d) — one entry per detected hand per sampled frame, not a frame count	`fourthuman_engine_api/sampledata/Egocentric_Sample_2_Annotation.json`	2026-09-11
185 frames `reference_run_kinematic_sampled_frames`	count of distinct tracks.kinematics_3d[].frame_idx	`fourthuman_engine_api/sampledata/Egocentric_Sample_2_Annotation.json`	2026-09-11
5 timeline tracks `reference_run_schema_tracks`	len(tracks)	`fourthuman_engine_api/sampledata/Egocentric_Sample_2_Annotation.json`	2026-09-11
413 HTTP routes `backend_routes`	regex count of router decorators in routers/*.py	`fourthuman_engine_api/routers`	2026-10-26
32 public pages `web_pages`	public routes in routes.json — the registry vite, the sitemap, the chrome and the SEO baseline all derive from	`fourthuman_web_suite/routes.json`	2026-10-26
6 domains `catalogue_capture_domains`	count of non-simulation capture briefs in the registry	`fourthuman_engine_api/config/capture_briefs.json`	2026-09-14
5 export formats `delivery_formats_implemented`	detected in fourthuman_engine_api/services/ by the writer or library each format requires — not asserted Implemented: FH-Ego v1 JSON, LeRobot v2.1, HDF5, Zarr v2, TFRecord (RLDS-style). MCAP and WebDataset are not offered.	`fourthuman_engine_api/services`	2026-10-26
47 durable stores `persistent_stores`	count of distinct .db filenames referenced in backend source Distinct stores, not connect() call sites. Store(s): accounts.db, address_proofs.db, api_keys.db, app_events.db, app_invites.db, app_uploads.db, audit_log.db, capture_briefs.db, capture_partner_profiles.db, clip_origin.db, cohorts.db, consent_ledger.db, consent_publication.db, console.db, crm.db, cycle_settlements.db, d1_copy.db, deliveries.db, earnings.db, engagements.db, engine_settings.db, equipment.db, founder_alerts.db, gear.db, head_mount.db, kit_forms.db, kit_mail.db, leads.db, messages.db, partner_ops.db, partner_sync.db, payout_requests.db, payouts.db, pipeline_jobs.db, push_devices.db, qa_verdicts.db, quotes.db, referrals.db, registrations.db, releases.db, request_throttle.db, retention.db, signoffs.db, tasks.db, transfer.db, web_sign_in.db, and one read-only archive file.	`fourthuman_engine_api/services`	2026-10-26
0 dev-only stores `in_memory_placeholder_stores`	count of PROVENANCE: placeholder markers Lost on process restart. No published figure may derive from these.	`fourthuman_engine_api`	2026-10-26

3 figures

Derived

Computed from values held in a committed artifact: an aggregate of a model's output, or a count that follows from a fixed rule. No new observation enters at this step.

Figure	How it is arrived at	Source artifact	Source last changed
16 segments `reference_run_action_segments`	len(actions): the exporter cuts the clip into fixed max(fps × 4, 60)-frame windows (120 frames at 30 fps, the last one shorter), so the count follows from the frame count rather than from boundaries found in the footage; the artifact lists actions[].start_frame and end_frame in provenance.derived_fields	`fourthuman_engine_api/sampledata/Egocentric_Sample_2_Annotation.json`	2026-09-11
0.8712 confidence (0-1) `reference_run_mean_hand_confidence`	arithmetic mean of tracks.kinematics_3d[].confidence — MediaPipe handedness classifier score, one per detected hand	`fourthuman_engine_api/sampledata/Egocentric_Sample_2_Annotation.json`	2026-09-11
0.508 confidence (0-1) `reference_run_min_hand_confidence`	min of tracks.kinematics_3d[].confidence — MediaPipe handedness classifier score, one per detected hand Published deliberately. The floor matters more than the mean for a buyer assessing worst-case quality.	`fourthuman_engine_api/sampledata/Egocentric_Sample_2_Annotation.json`	2026-09-11

3 figures

A real zero

Each is a real, deliberate zero, published as a figure.

Figure	How it is arrived at	Source artifact	Source last changed
0 Commissioning labs `named_customers`	count of Commissioning labs that have commissioned a Capture Brief	no artifact — see the method	—
0 attestations `third_party_attestations`	count of third-party security attestations held	no artifact — see the method	—
0 hours `captured_dataset_hours`	hours recorded by Capture Partners against a commissioned Capture Brief Catalogue volumes are published as target figures.	no artifact — see the method	—

2 figures

Target

What a brief is scoped to produce, never what exists. A target may not be read as inventory or as capability.

Figure	How it is arrived at	Source artifact	Source last changed
182,700 clips (target) `catalogue_target_clips`	sum of target_volume across non-simulation capture briefs — every input is itself a target agreed per brief, never delivered inventory	`fourthuman_engine_api/config/capture_briefs.json`	2026-09-14
1,505 hrs (target) `catalogue_target_hours`	same sum, hours. Excludes simulation-traversals: 10,000 of those hours are simulated scenes, a different product that is not addable to human capture	`fourthuman_engine_api/config/capture_briefs.json`	2026-09-14

Recompute them yourself

The table above is generated from a single file, and that file is generated from the artifacts named in it. One command rebuilds the lot:

python3 .agents/generate_facts.py

If a figure on this page disagreed with the artifact it names, that command would change the file and our build would fail.

This page lists the figures generated into facts.json. A number written straight into a sentence on another page is not generated, and is not listed here.

FH-Ego v1, field by field · The clip most of these are computed from