Egocentric Hotel Amenity Restocking Data

Egocentric hotel amenity restocking data is first-person video of one person restocking the items a guest room carries, recorded from that person’s own point of view while they work. A paid, consented contributor wears the capture rig and performs the task written in a skill spec, and each attempt is delivered as an episode with synchronized streams, hand pose and labeled action segments. It is human capture. It is not robot or teleoperation data, and it is not a fixed corpus you download.

The site states the task in one row of its commercial task catalog. In the Hotel & Lodging environment, the row "Hotel Amenity Restocking" reads: "Restock toiletries, towels, coffee and minibar items from the housekeeping cart." It is an Easy row, captured with a head-mounted phone or a multiview rig, and listed as ready to deploy. The environment itself is described with examples of hotel guest room servicing, resort housekeeping, bed-making, linen changes, bathroom cleaning and amenity restocking. This page is built on that row and on the neighbouring rows quoted below, and on nothing else the site says about hotels.

Neighbouring pages cover different subjects. Egocentric hotel room turnover data covers the whole reset of a room, egocentric bathroom cleaning data centres on the bathroom, and egocentric napkin and towel folding data centres on folding. The subject here is the restock itself: taking items from the cart and putting them where they belong.

People reach this topic through hotel amenity restocking robot training data, egocentric minibar and toiletry restock video, a housekeeping cart dataset, or custom hospitality data collection. They describe one request from different angles. The work is the same: write the task down, collect it on demand, and receive validated episodes.

Three catalog rows name restocking, and the table below reproduces them as the catalog states them. "Hotel Amenity Restocking" is the lead row: Easy, Hotel & Lodging, captured with a phone (head-mounted) or a multiview rig, Ready to deploy. "Clean and restock guest bathrooms using disinfectant and amenity kits" is also Easy and Ready to deploy in Hotel & Lodging, with the same capture setup; its description reads "Clean and disinfect the guest bathroom and restock towels and amenity kits." "Restock amenities" sits in the Hospitality Service Set and reads "Restock a minibar/amenity tray." It is Easy, uses a multiview rig with handheld grippers, and is Available on request.

The three rows are not interchangeable. The first two are hotel rows and share a capture setup and a status. The third is a staged working set, described as a table, a counter or bar, and a housekeeping or room area, recorded by a multiview rig while the operator uses handheld grippers so that demonstrations map onto a robot end effector. A request that says only "amenity restocking" leaves open which of the three is meant, so the spec names the row.

The site names no amenity brand, no item count, no cart model and no minibar layout. The words that appear are toiletries, towels, coffee, minibar items and amenity kits, and this page uses only those. Whether a coffee item means a capsule, a sachet or something else, and what counts as a toiletry, is not stated. Those choices belong in the spec, and a Spec Pilot is the place to settle them.

A broader row, "Hospitality Housekeeping", also mentions restocking: "Strip and make beds, clean bathrooms, dust and vacuum and restock guest rooms." It is Easy and Ready to deploy. It is the row to look at when restocking is one step in a longer pass rather than the subject of the episode.

Each attempt is delivered as one episode, and a restock needs a decision up front about what one attempt is. It could be a single placement, one item type, one location such as a minibar or a bathroom shelf, or the full run from the cart to every place the cart’s items go. The site does not set an episode length, a number of items per episode or a number of episodes, so the skill spec does. Inside the episode, the synchronized streams run from start to end, and action segments divide it into labeled spans.

A short episode is easier to check against a single accept criterion, such as one towel placed where it belongs. A long episode shows the order of the work: which items come off the cart first, where the contributor moves next, and when the restock is finished. That is an illustration of how a spec can cut an episode, not a stated duration. Many programs mix the two, with short episodes for the steps a model is weakest at and long episodes for the sequence.

The cut matters because of one machine-checked rule. An episode is rejected if the manipulated object is out of frame for more than 15 percent of it. In a restock the manipulated object changes from a towel to a toiletry to a coffee item, and the contributor turns between the cart and the place the item goes. The spec should say what counts as the manipulated object at each step so that the rule can be applied the same way across contributors.

The catalog row says "from the housekeeping cart", which means the cart is the starting point of each reach. Whether the cart is in frame throughout, and whether it is parked at the door or inside the room, is not stated. Ask for it in the spec if your model needs to see the source and the destination in one view.

Inside each episode, action segments divide the recording into start- and end-timestamped spans, each carrying start_ns, end_ns, verb and noun. For this task family a segment can be one step as written in the skill spec, so a restock becomes a run of labeled steps such as take, carry, place and replace, with a noun for the item, rather than one long clip. The other standard fields are hand_joints and hand_visible for 3D hand pose, contact_events for the moment the hand meets an object, objects, and instance_masks, stored at the depth resolution of 848×480 so they line up with the depth stream.

Hand pose is delivered for both hands with a per-joint visibility flag for occluded fingers. That is the label that carries how the hands pick a small item off a cart, hold it and set it down, and it is the main signal for the manipulation inside a restock. The site describes contact events as the moment the hand meets an item, and it does not say which items count in a hotel room, so the spec is the place to settle what a contact event means for your task.

The site lists no label for the outcome of the restock. There is no stated stock level, item-present check, minibar-state label, par level or task-completion score, and no contact force reading. The catalog verb "restock" describes what the contributor does, not a delivered measurement. Camera pose, a head trajectory and a map of the room are likewise not among the listed deliverables. The site states 6-DoF IMU data at 200 Hz from head- and wrist-mounted sensors, and nothing more about motion.

Delivery is in the format your stack already reads: RLDS, LeRobot, WebDataset, HDF5, zarr or Rerun .rrd, or raw streams with metadata. Readable converter source ships alongside the data so the schema can be retargeted to your own loader. The official documentation for three of these formats is linked in the sources below. Collections are billed per validated hour, and the license is perpetual, irrevocable and worldwide, buyer-owned and non-exclusive by default, with exclusivity available as a separate term.

A published robot foundation model paper, GR00T N1, linked in the sources below, describes a data pyramid with web data and human video at the base, synthetic data in the middle and real robot trajectories at the top. First-person human video of a restock sits in that base layer. Human capture does not replace robot data, and this page does not claim that it does. The glossary entry on the data pyramid explains how the layers trade volume for embodiment closeness.

Restocking is a useful shape of task because it is made of small, repeated pick-and-place moves on items that differ in size and softness. A towel deforms and a toiletry does not. A coffee item and a minibar item are different again. The catalog groups them in one row, so a single episode can contain several object types, each needing a different grasp. A clip that shows a cart somewhere in frame does not carry what a policy needs. The hands, the items and the step boundaries have to be in the episode.

A custom collection lets you aim at the step your model is missing. If your pipeline already handles the pick and fails at the placement, or fails at the sequence, the spec can name that. This page does not claim that any public corpus lacks restocking data, and it does not claim a training outcome or a model improvement from any dataset. It also does not claim an off-the-shelf restocking dataset; this is custom collection.

The tradeoff to plan for is the embodiment gap between a human pair of hands working from a cart and a robot platform doing the same, which retargeting or pairing with robot data has to close. Counts of contributors, hours and tasks are set in your skill spec.

The site describes the two capture setups in the catalog. For the two hotel rows, a head-mounted phone records the first-person view of real work on site, and a multiview rig adds synchronized third-person angles when the spec asks for them. For the service set row, the setup is a staged working set recorded by a multiview rig while the operator performs the task with handheld grippers. "Ready to deploy" means that contributors and site access are set up, and capture starts once a spec is signed. "Available on request" is a different status, and the service set row carries it.

The site does not describe stereo capture for any of these three rows. A calibrated stereo pair is offered generally, and if your program wants one for restocking, it is agreed in the skill spec. The stereo hotel page covers that offer for four other hotel rows. This page makes no stereo claim of its own.

The site names no property, brand, city, room count or room type, and it lists capture at your own facility as not offered. This page therefore makes no claim that episodes are recorded in a particular hotel, and none that they are recorded at yours. Capture is run to your skill spec using the settings and contributors Firsthand arranges. The hospitality page classes back-of-house work such as room turnover separately from front-of-house settings, which it states are bystander-dense with lower yield after de-identification.

Every contributor signs a written release before capture, and each delivered episode links to that record through consent artefact IDs in the batch data card. At ingest, automated detection finds faces, plates and screens and blurs them irreversibly, and speech is scrubbed of names, addresses and card numbers while ambient audio is kept. A human reviewer confirms that no un-consented person is identifiable in any frame, and footage that cannot be de-identified is discarded. An undocumented bystander is also one of the machine-checked reject criteria. The anonymization record is delivered with the data. The site states nothing specific about guest belongings or room contents, so this page makes no claim beyond the general process.

On chemicals and safety, one row names "disinfectant" in its title. The site states no chemical-handling protocol, no contributor safety protocol, no hygiene or infection-control guidance and no hospitality certification for any restocking row, and this page does not describe one. If your program needs terms about the materials used or the conditions under which the work is done, raise them when you scope it. None of this is legal advice, and the executed agreement controls.

A restocking collection uses the same capture stack as every Firsthand program. The instrumented tier is a head-mounted stereo rig that records first-person RGB, with a metric depth stream recovered from the calibrated stereo rig rather than from a standalone depth sensor, and 6-DoF IMU data at 200 Hz from head- and wrist-mounted sensors. The wider stack is ten streams: a 4K head camera, two wrist cameras, depth, IMU, 3D hand and body pose, instance segmentation and action labels, hardware-aligned under 2 ms. The two wrist cameras give a closer view of the hands than the head camera alone, which matters when a small item is held in a closed hand.

Every stream is hardware-triggered onto one PTP clock domain, so depth, motion and hand pose line up with the same video frame rather than being approximated afterwards. The rig is calibrated twice per session, before capture starts and after it ends, and an episode is accepted only if RMS reprojection stays under 0.28 px. The site also says depth degrades on wet and reflective surfaces and handles that with per-frame confidence maps. A guest bathroom has mirrors, fixtures and wet surfaces, but the site does not say how depth behaves on any of them, so this page makes no accuracy claim for them.

The site does not state a camera count, resolution or frame rate specific to restocking, and it names no device, cart or tool beyond the housekeeping cart and the handheld grippers of the service set. Force and tactile readings are not among the streams, and no robot or teleoperation data is captured. Which tier applies to your program is set in the skill spec.

The table below sets this page against the neighbouring hotel and hospitality pages on this site. Each row states the page’s main subject and the catalog rows it draws on, with the status as the catalog lists it. No delivered hours are claimed for any of them.

The short version: the room turnover pages treat the reset of a whole room as one task family, the stereo hotel page covers the stereo and container side of four turnover rows, the lobby page covers public areas, the bathroom page covers cleaning and its related restock row, and the folding page covers folding. This page treats the restock itself as the subject, with the cart as the source.

From signed spec to first validated batch, the site states a median of 48 hours where coverage is deep and three to four weeks where it is thin. The site names hospitality among its thin coverage areas, where thin environments typically take three to four weeks, and it does not say whether the Hotel & Lodging environment counts as deep, so ask which timeline applies when you request the quote. Speed comes from a warm pipeline, not from cutting consent, review or anonymization.

The site also states program milestones: a pilot reaches its first delivery in two to three weeks, and a standing program reaches full rate in four to six weeks. Those assume capture coverage is already in place. After the first batch, on-demand top-ups run against the same accept criteria, so you can add items or locations without rewriting the spec. The site does not state a turnaround for those later batches, so plan them in your quote.

Only footage that clears the signed accept criteria is billed as a validated hour, and rejected items are recaptured at our cost. The site does not publish a price list, a minimum order or a batch size for this task family, so those are set in your quote. Collection is fast and on demand in the sense that you do not need a standing program: a focused brief is a normal starting point.

Where this site does not state something, this page says "Not stated" rather than filling the gap. There is no amenity brand, no item or toiletry list beyond the words in the catalog rows, no cart specification, no minibar layout, no property or room type, no task duration, no site count, no region or city, no episode count, no stock-level or completion label, and no chemical-safety guidance. Nothing here reports a robot result, a benchmark or a model improvement.

Counts of contributors, hours and tasks are set in your skill spec. A task outside the rows quoted above, such as a restock routine the catalog does not name, is settled in a Spec Pilot rather than assumed.

Describe the items, the places they go, the row you mean and the variation or failure cases your pipeline is missing. That becomes a written skill spec, quoted for reach, timeline and price before any capture begins, and most programs run a Spec Pilot first, a small, fast first batch that proves the spec and the accept criteria on real footage before volume scales. The scoping checklist below maps the parts of a skill spec onto a restocking request. A published sample recording is on the sample data outputs page if you want to see the streams before you write a spec.

Three choices matter most for this family. Name the item and the destination rather than a broad activity, because that is what lets each placement become one action segment. Decide which catalog row you mean, since the hotel rows and the service set row differ in capture setup and status. And name where the episode starts and stops: one placement, one location or the whole run from the cart, since the site does not set an episode length.