A named three-step sequence
Collecting from the nests, checking against a light and packing are different steps, so a spec names each one and where it runs.
Task family
Requirements
Collecting from the nests, checking against a light and packing are different steps, so a spec names each one and where it runs.
The catalog puts the row in Farm & Field Operations without naming the bird species or the type of farm, so the spec states which setting it covers.
A head-mounted view or a multiview rig keeps the hands, the eggs, the light and the packing container in frame while the work happens.
Frame-accurate action segments mark where collecting ends, where the check against the light starts and where packing begins.
A perpetual, buyer-owned commercial license backed by a documented consent chain.
Why it matters
Egocentric egg collection and candling data is real-world video of the physical hand work of gathering eggs and getting them ready to pack: collecting from the nests, checking each egg against a light for defects and packing the ones that go forward. It is recorded against a written spec, with the streams the site states for its custom collections.
The task has one catalog row, and until now it appeared only as a line in a longer list. The farm and field page lists it among the Farm & Field Operations rows, and the livestock page names it as an adjacent row that the catalog does not call livestock handling. This page treats the row as its own subject. It describes task video only, and it makes no food-safety, grading-accuracy, animal-welfare or certification claim.
Catalog row
One row. "Egg Collection & Candling" sits in the Farm & Field Operations environment under the Agriculture & Grounds sector. The catalog describes it as: collect eggs from the nests, check them against a light for defects and pack them. It is rated Easy and listed as Ready to deploy, and its capture mode is a head-mounted phone or a multiview rig.
The catalog gives one line for the row. The steps inside that line, and what counts as a defect in your pipeline, are set by your spec.
Nearby rows
Three nearby Farm & Field Operations rows are separate listings. "Feeding & Watering Rounds" and "Feed and water livestock using feed buckets and hoses" cover feed and water for animals. "Animal Pen Cleaning & Waste Removal" covers mucking out pens. "Crop Sorting & Grading" is a different row: sort harvested crops by size and quality, discard damaged items and grade into crates. It is a crop row, and the site does not describe it as egg work.
The site lists no row for egg washing, egg grading by weight, incubation or hatchery work, or processing eggs in a packing plant. Ask for any of them in the spec.
Streams
Firsthand custom collections record ten hardware-synchronized streams aligned under 2 ms, with RGB, depth, 3D hand pose with a per-joint visibility flag, contact and grasp events and frame-accurate action segments. The one language label the site states is a task instruction for each episode.
Standard capture is cameras, depth and IMU. A separate glove tier records a tactile signal, and the site does not state whether it applies to this row. The streams that carry force or tactile data are therefore settled in the spec.
Privacy
This page does not describe capturing farm owners, staff or visitors beyond the contributor doing the task. Every collection runs on a documented consent chain. Faces are irreversibly blurred at ingest, a reviewer confirms the result, and footage with bystanders who cannot be de-identified is discarded.
Fast · on demand
Where coverage is already deep, the first validated batch lands in a median of 48 hours from signed spec. Thin environments typically take three to four weeks, because vetting operators there is the bottleneck, and we quote that up front. The site does not state which of the two applies to this row.
You describe the sequence, the setting and the birds or props your policy needs to see, and we turn that into a written spec covering tasks, environments, capture tier, annotation depth and accept criteria. A Spec Pilot, a small and fast first batch, can prove the spec before you scale. Billing is per validated hour, a reject log ships with every batch, and rejected items are recaptured at our cost.
At a glance
Every collection runs through the same seven-stage end-to-end custom collection pipeline. End-to-end custom data collection is a managed service that takes an AI data need from problem to owned dataset in a single accountable pipeline.
Catalog row
| Catalog task | Environment and sector | Catalog description | Skill | Status |
|---|---|---|---|---|
| Egg Collection & Candling | Farm & Field Operations, Agriculture & Grounds | Collect eggs from the nests, check them against a light for defects and pack them | Easy | Ready to deploy |
| Feeding & Watering Rounds (separate row, not egg work) | Farm & Field Operations, Agriculture & Grounds | Walk the pens or fields delivering feed and topping up water for the animals | Easy | Ready to deploy |
| Animal Pen Cleaning & Waste Removal (separate row, not egg work) | Farm & Field Operations, Agriculture & Grounds | Muck out pens with fork and shovel, remove waste in barrows and lay fresh bedding | Easy | Ready to deploy |
| Crop Sorting & Grading (separate row, crops) | Farm & Field Operations, Agriculture & Grounds | Sort harvested crops by size and quality, discard damaged items and grade into crates | Easy | Ready to deploy |
Streams and tasks
| Stream or label | Status on this site | What it carries in an egg collection episode |
|---|---|---|
| RGB and depth | Stated; ten streams synced under 2 ms | Keeps the nest, the eggs, the light and the hands on one timeline |
| 3D hand pose, per-joint visibility | Stated for custom collections | Carries how fingers pick up, hold and set down an egg |
| Contact and grasp events | Stated | Marks when hands meet an egg, the light or the packing container |
| Action segments | Stated; frame-accurate | Separates collecting, checking against the light and packing |
| Task instruction | Stated; one per episode | States the egg collection task in plain language |
| Object labels | Set by your spec | Names the items and steps in the terms your policy uses |
What is not stated
| Item | Status on this site | Where it is settled |
|---|---|---|
| Bird species, flock or housing type | Not stated | The written spec |
| Whether birds are present during capture | Not stated | The written spec |
| What counts as a defect | Not stated | The written spec |
| Food-safety, grading-accuracy or animal-welfare claims | Not stated | Not claimed on this page |
| Certification or regulatory claims | Not stated | Not claimed on this page |
| Candling lamp or equipment models | Not stated; the catalog names item types only | The written spec |
| Farm names, regions, number of sites | Not stated beyond Western program markets | The written spec |
| Cycle times, throughput, eggs per episode | Not stated | Quoted against your spec |
| Number of episodes or workers | Not stated | The written spec |
| Success or quality labels | Not stated | The written spec, if needed |
| Force or tactile data | Not stated for this row | The written spec, if needed |
| Robot or policy results | Not stated | Not claimed on this page |
Formats
| Format | What the official documentation states | Used for here |
|---|---|---|
| RLDS | A dataset is retrieved as episodes, and each episode holds a dataset of steps (RLDS repository README) | Episode-based training loaders |
| LeRobot | LeRobotDataset v3.0 is a standardized format for robot learning data, including multi-camera video, with many episodes stored per Parquet or MP4 file (LeRobot documentation) | LeRobot dataset format and tooling |
| WebDataset | Tar files in which files that belong together share a basename, split into numbered shards (WebDataset README) | Sharded streaming for large training runs |
| HDF5 and zarr | Stated by Firsthand only | Array storage for custom pipelines |
| Rerun .rrd | Stated by Firsthand only | Multi-stream synchronized playback for review |
FAQ
Sources
Tell us the egg collection sequence, the farm setting and the birds or props your policy needs to handle. We will quote reach, timeline, and price against a written spec, with a first validated batch in a median of 48 hours where coverage is deep.