Both hands and the piece in frame
A head-mounted view keeps both hands, the small metal piece and the tool in one view while the joint is made, the piece is polished or the wire is wrapped.
Now recruiting capture teams worldwide. Get paid to record first-person video of real work.
Apply to captureTask family
Requirements
A head-mounted view keeps both hands, the small metal piece and the tool in one view while the joint is made, the piece is polished or the wire is wrapped.
3D hand pose with a per-joint visibility flag records where the fingers are, and which joints were hidden, while the work is held close.
Frame-accurate action segments mark where fluxing, heating, wrapping or polishing ends and the next step begins, aligned to video and depth.
A perpetual, buyer-owned commercial license backed by a documented consent chain.
Why it matters
Egocentric jewelry soldering and polishing data is real-world, first-person video of the joining, forming and finishing work at a jeweler bench. It is recorded from the worker viewpoint, with depth, 3D hand pose, contact events and frame-accurate action segments aligned on one hardware clock. It covers three tasks the site states in its task catalog: Soldering, Polishing and Wire wrapping.
The broader jewelry and metalsmithing page covers all five bench rows, including Engraving and Gem setting. This page narrows to the three rows about joining, finishing and forming, so a team that needs egocentric jewelry soldering data, first-person metal polishing video or a wire wrapping jewellery dataset can scope exactly those tasks. It is custom bench jewelry data collection, fast and on demand, written to your spec.
Catalog rows
The site lists them in a Jewelry & Metalsmithing Studio environment under the Creative Workshops & Crafts sector. The rows are Soldering ("Flux the joint, heat with a torch and flow solder to join metal pieces.", Hard), Polishing ("Polish metal pieces on the wheel with compounds and finish by hand.", Medium) and Wire wrapping ("Bend and wrap wire around stones and beads with pliers to form jewellery.", Medium).
All three list the capture setup "Phone (head-mounted) / multiview rig" and the status Available on request. The environment lists its examples as goldsmith bench, gem setting, wire wrapping, soldering, engraving, polishing. The first table below reproduces the three rows as the catalog states them.
Soldering
The catalog row reads "Flux the joint, heat with a torch and flow solder to join metal pieces." It runs from fluxing the joint through heating with a torch to flowing the solder, so one row covers the whole joining sequence at the bench. It is the only row on this page rated Hard.
The site names a torch and no more. Torch type, solder type, temperatures and safety practice are Not stated, so this page makes no claim about them and gives no safety guidance. They are set in the written spec. Electronics soldering is a different task in a different environment and is covered by its own page.
Polishing
The catalog row reads "Polish metal pieces on the wheel with compounds and finish by hand." It has two parts, machine polishing on a wheel with compounds and then a hand finish, and it is rated Medium.
Wheel models, compound types and the finish a piece is meant to reach are Not stated, and the site states no outcome label for how a piece looks afterwards. Polishing in other settings, such as car detailing, is a different task covered elsewhere on the site and is not counted here.
Wire wrapping
The catalog row reads "Bend and wrap wire around stones and beads with pliers to form jewellery." It covers bending and wrapping wire by hand and with pliers, around stones and beads, and it is rated Medium.
Wire gauges, stone and bead types and the finished designs are Not stated, so the spec sets which pieces a session makes. A wire wrapping jewellery dataset is therefore scoped by your written description of the pieces and the steps, not by a fixed catalog list.
Compared with neighbouring pages
Four neighbouring pages exist, and they are not the same topic. The jewelry and metalsmithing page covers all five rows as a task family, the gem setting and engraving page covers the two cutting and stone rows, the stereo jewelry page covers the stereo and MCAP side of all five rows, and the soldering and PCB rework page covers electronics soldering.
This page is the narrow view for teams scoping joining, finishing and forming only. The second table below sets the four against this page. Where a team needs Engraving or Gem setting as well, the broader jewelry page is the better starting point.
Capture setup
Not as the site states it. The catalog lists "Phone (head-mounted) / multiview rig" for every row on this page and states no stereo pair for them. The site does offer a stereo pair through its general stereo offer, so stereo is a capture choice agreed in the written spec and proved on a Spec Pilot, not something these rows already run. The stereo jewelry page covers that side.
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. The site states a head camera at 3840 × 2160 and 60 fps for its instrumented rig, but no camera count, resolution or frame rate for these rows, so the capture tier is agreed in the spec. Throughput is Not stated and is quoted against your spec.
Why it is hard
The rows describe work with a torch, compounds, a wheel and pliers on small metal pieces, finished by hand. The site does not state piece sizes or tolerances, so this page makes no claim about how small the work is. What the site does state is what egocentric video puts into the training distribution: hands entering the frame from below, occlusion during grasp and motion blur during reach.
The site also states why a hand pose stream matters here. RGB alone leaves a model to infer 3D hand structure it cannot see reliably through occlusion and motion blur. Firsthand delivers 3D hand pose as handJoints[2][21][3] in metric units in the head-camera frame, with a visibility flag per joint, aligned to depth and RGB. Two hands of 21 joints each are tracked per frame, and the flag records which of them were hidden.
Streams and labels
Each episode follows the Firsthand episode schema: ten streams synced under 2 ms, with head and wrist RGB, depth, 3D hand pose with per-joint visibility, contact and grasp events, and frame-accurate action segments. For these rows, hand pose tracks how the fingers hold and move the piece or the wire, contact events mark when fingers or a tool meet it, and action segments mark each step, such as fluxing, heating, wrapping and polishing.
The only language data the site ties to Firsthand capture is one task instruction per episode. Object labels use the terms your spec sets. The data is human capture by paid, consented contributors, not robot or teleoperation data. Force or tactile data, and success or quality labels for a finished piece, are Not stated for these rows, and this page does not offer them as deliverables.
Gloves
The site offers a tactile glove tier as a pilot-ready option. The gloves track the hands, the glove signal is time-synced to a head-mounted stereo camera by hardware timestamps and an IMU, each glove is calibrated every session, and the pipeline checks signal dropout and sync drift.
The site does not state that the glove tier has been used on any jewelry row, and it does not state whether gloves fit bench work with a torch, a wheel or fine wire. Treat force and tactile data as Not stated for these rows unless a spec adds the glove tier and agrees the fit with you. The tactile glove page covers the tier in full.
Setting and privacy
The catalog places all three rows in the Jewelry & Metalsmithing Studio environment. The site names no studio, city, school or jeweler, and it does not state capture at a customer own studio or facility. Environment, access and schedule are agreed in the spec, and the number of studios is Not stated.
Every episode links to a signed contributor release through a documented consent chain. Faces, plates and screens are detected and irreversibly blurred at ingest, a reviewer confirms the result before anything ships, and the anonymization record is delivered with the data. An undocumented bystander is also a stated reject criterion, and the site states no further privacy rule specific to studios.
Quality gates
Episodes are machine-checked at ingest against the stated reject criteria, listed in the fourth table below, and a human reviewer then checks the result. The site states them for every episode, so they apply to soldering, polishing and wire wrapping alike.
It does not state how often a small piece leaves the view at a bench, and it states no blur or sharpness rule. A rejected item goes to the reject log and is recaptured at our cost.
Delivery
Delivery formats are RLDS, LeRobot and WebDataset, plus HDF5, zarr and Rerun .rrd. The last table below sets each against what its official documentation states. These episodes use the same schema as every other Firsthand collection, so a loader that reads another Firsthand dataset reads this one without changes.
Every batch is validated against signed accept criteria before it ships, with a reject log, and delivery is rolling, per validated batch.
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.
On-demand top-ups run against the same spec and accept criteria after each training run. All three rows are Available on request, so the site states no capture timeline for them beyond a quote against your spec.
The spec
You describe the pieces, steps and sequence your policy needs: where a joint starts and ends, what a polished piece is, how many wraps make a finished item. We turn that into a written spec covering tasks, environments, capture tier, annotation depth and accept criteria, then quote reach, timeline and price before any capture begins.
A Spec Pilot, a small and fast first batch, can prove the spec and accept criteria before you scale. Billing is per validated hour, a reject log ships with every batch, rejected items are recaptured at our cost, and every collection ships under a perpetual, buyer-owned commercial license.
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 rows
| Task (as listed) | Description (as listed) | Environment | Skill | Capture setup | Status |
|---|---|---|---|---|---|
| Soldering | Flux the joint, heat with a torch and flow solder to join metal pieces. | Jewelry & Metalsmithing Studio | Hard | Phone (head-mounted) / multiview rig | Available on request |
| Polishing | Polish metal pieces on the wheel with compounds and finish by hand. | Jewelry & Metalsmithing Studio | Medium | Phone (head-mounted) / multiview rig | Available on request |
| Wire wrapping | Bend and wrap wire around stones and beads with pliers to form jewellery. | Jewelry & Metalsmithing Studio | Medium | Phone (head-mounted) / multiview rig | Available on request |
Comparison
| Page | Main topic | Rows or tasks covered | How it differs from this page |
|---|---|---|---|
| This page | Jewelry soldering, polishing and wire wrapping | Soldering, Polishing, Wire wrapping | The narrow view for joining, forming and finishing |
| Egocentric jewelry and metalsmithing data | The whole jewelry bench as a task family | All five rows, including Engraving and Gem setting | Broader; start here if you need all five rows |
| Egocentric gem setting and engraving data | Seating a stone and cutting metal | Gem setting and Engraving, both Hard | Covers none of the three rows on this page |
| Stereo jewelry and metalsmithing data | The stereo and MCAP side of jewelry work | All five rows | A stereo sibling; the site states no stereo pair for these rows, so stereo comes through the general offer |
| Egocentric soldering and PCB rework data | Electronics soldering at a board | The electronics soldering rows | A different environment and task, not jewelry |
Streams and tasks
| Stream or label | Status on this site | What it carries in these episodes |
|---|---|---|
| Head and wrist RGB, depth | Stated; ten streams synced under 2 ms | Keeps the piece, the tool and both hands on one timeline |
| 3D hand pose, per-joint visibility | Stated; two hands of 21 joints, with a visibility flag per joint | Carries how the fingers hold the piece or wire and which joints were hidden |
| Contact and grasp events | Stated | Marks when fingers or a tool meet the piece |
| Action segments | Stated; frame-accurate | Separates fluxing, heating, wrapping and polishing |
| Task instruction | Stated; one per episode | Names the task the episode covers |
| Object labels | Set by your spec | Names the pieces and steps in the terms your policy uses |
| Force or tactile data | Glove tier stated; use on jewelry rows Not stated | Added only if a spec includes the glove tier |
Quality gates
| Reject criterion | Status on this site | At a jewelry bench |
|---|---|---|
| Manipulated object out of frame for more than 15 percent of the episode | Stated | Applies to every episode; how often a small piece leaves the view is Not stated |
| More than 35 percent invalid depth pixels in the manipulation region | Stated | Applies to every episode; depth behaviour on metal is Not stated |
| Sync breach beyond the ceiling for more than three consecutive frames | Stated | Applies to every episode |
| Calibration drift beyond tolerance | Stated | Applies to every episode |
| Undocumented bystander | Stated | Applies to every episode; the episode is rejected |
| Mismatch against the named skill spec | Stated | Checked against the task and steps your spec names |
| Blur or sharpness rule | Not stated | Set in the spec if your policy needs one |
What is not stated
| Item | Status on this site | Where it is settled |
|---|---|---|
| Metal types, stone types, bead types, wire gauges | Not stated; rows name metal pieces, stones, beads and wire only | The written spec |
| Torch, solder, wheel and compound types | Not stated beyond the torch, wheel, compounds and pliers named in the catalog rows | The written spec |
| Temperatures, tolerances and safety practice | Not stated | The written spec |
| Piece sizes and number of pieces or episodes | Not stated | The written spec |
| Number of studios or jewelers, skill levels | Not stated | The written spec |
| Stereo pair on these rows | Not stated for the rows; offered generally | The written spec, proved on a Spec Pilot |
| Throughput and capture timeline | Not stated | Quoted against your spec |
| Force or tactile data on jewelry rows | Not stated | The written spec, if needed |
| Success or quality labels, robot results | Not stated | The written spec, if needed |
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 |
Example footage
One real clip already published on this site, shown as a format and rig example: ten seconds of first-person home cooking with faces blurred. It does not show soldering, polishing or wire wrapping.
No jewelry clip is published on the site, so this is labelled as a format example. Request a sample to see delivered files. See sample data outputs, request an egocentric data sample.
FAQ
Sources
Tell us the pieces, steps and sequence 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.