voice hardware

Card, Clip or Pendant: Choosing a Voice Wearable

Compare card, clip and pendant voice recorder forms by mic position, wearing habit, button use, battery, weight and what triggers full ODM tooling.

Product photo: Hearit.ai HA-MIC06 (AI VoiceRec).
Product photo: Hearit.ai HA-MIC06 (AI VoiceRec).

The form factor is already part of your audio stack.

For a voice product, card, clip and pendant are not cosmetic choices. They decide where the microphone sits, how often the user touches the device, how much battery you can carry and whether your request fits an existing platform or becomes new hardware.

Use the Mouth-Mount-Moment-Mass model.

Mouth: mic position. Mount: how it is worn. Moment: what the button does. Mass: weight, battery and comfort.

Mouth: where the microphone actually hears from

A card-style recorder, especially one attached to the back of a phone or carried flat, is convenient but inconsistent. The microphone may face a pocket, table, palm or jacket depending on how the phone is handled.

A clip-on recorder is usually more predictable because it can sit near the collar, lapel or bag strap. A pendant can also sit near the chest, but it moves as the user walks, bends or turns.

  • Card: good for phone-adjacent workflows, less stable for microphone direction.
  • Clip: strong for chest-level pickup when users accept clipping it in place.
  • Pendant: easy to wear, but swing and fabric contact need attention.

If your app depends on short, deliberate recordings, a card may work. If your product depends on workday capture, note-taking or field conversations, mic placement becomes a product requirement, not a preference.

Mount: all-day wear changes behavior

Card-style devices win when the phone is already the center of the workflow. They are easy to remember because the user already carries the phone, but magnetic placement, case thickness and pocket orientation can affect real use.

Clip-on devices ask the user to choose a location every morning. That adds friction, but it can also create repeatable audio if the clip is simple and the device is light enough to forget.

  • Pendant habit: easiest to put on, especially for roles that already use badges or lanyards.
  • Clip habit: best when clothing or uniform placement is predictable.
  • Card habit: best when the phone is always present and the recording session is intentional.

This is where many teams should start with a platform rather than new tooling. GMIC’s existing BLE and Wi-Fi voice platforms can support white-label and OEM device options where the customer keeps their own app, speech-to-text, AI and data.


Moment: the button is part of the UX

Push-to-talk, press-to-bookmark, press-to-start and slide-to-control are different products. The button needs to match the recording model before anyone argues about enclosure shape.

On a card, the user may press while looking at the phone. On a clip or pendant, the user often presses by feel, which makes tactile feedback and accidental activation more important.

The right wearable is the one a user can operate without changing the conversation.

Button logic, recording behavior, upload endpoint, device identity and SDK integration can often stay within a platform customization path. GMIC’s SDK covers pairing and reconnection, recording control, file transfer and sync with retry, device status, OTA firmware updates and authentication to the customer endpoint.

Mass: battery, weight and patience

Battery is not separate from form. A thin card pushes you toward a flatter battery and tighter mechanical packaging, while a pendant or thicker clip may allow a different battery volume and antenna placement.

Existing GMIC platforms give useful reference points. MIC01 is a push-to-talk Bluetooth 5.3 platform with 128 MB memory, up to 8 hours of battery life and an 8.8 g weight; MIC05 and MIC06 add Wi-Fi plus Bluetooth 5.3, 1 GB memory and direct cloud or private-server paths, with higher weight and battery capacity.

  • Lighter devices: easier to wear, but less room for battery and mechanical features.
  • Heavier devices: may support more capability, but the mount must control bounce, pull and fatigue.
  • Wi-Fi upload: useful when the device should send audio to the customer cloud endpoint without depending only on phone sync.

This is why the future of voice hardware is moving away from one generic recorder shape. AI apps are becoming more vertical, and the winning device form will be the one that matches a specific workflow before competitors train users around theirs.

What stays on an existing platform?

White label is the lightest path: logo, packaging, manuals and app branding. Platform customization or OEM can change firmware behavior, button logic, upload endpoint and SDK integration without changing industrial or mechanical design.

That path is typically the right place for an MVP if you can accept the existing enclosure and prove the recording workflow first. Around 500 units is the reference point for white label and platform customization, and proof-of-concept runs on an existing platform are evaluated case by case.

Before asking for a new shape, test the recording moment.

A platform-based prototype or MVP typically takes 25–45 working days after scope and specifications are agreed; component supply is the main variable.

What makes it a full ODM program?

A true card, clip or pendant redesign usually means new ID, mechanical design, tooling, PCB layout, audio path and battery planning. Magnets, clip mechanisms, lanyard structures, microphone port direction and enclosure thickness can all push the work into a full ODM program.

Full ODM is also the path if you want to add display, camera or cellular connectivity. Those features change the product architecture, not just the shell.

  • White label: fastest when the existing device shape is acceptable.
  • Platform OEM: right when behavior and cloud integration matter more than new mechanics.
  • Full ODM: right when the form is a core reason users will adopt the product.

When GMIC is not the right fit

GMIC is not the right fit if you need a finished consumer gadget with a shared app and shared AI stack. The customer keeps their own app, speech-to-text, AI and data on every path.

It is also not the right fit if you need a one-off enclosure experiment with no path to production. Full ODM programs are planned around new tooling and roughly 2,000 units, while existing-platform paths are better for earlier validation.


If you are comparing card, clip and pendant, start by deciding whether form is a workflow test or a product moat. Budget ranges and a plan come from a 30-minute call, but the first fork is clear: prove behavior on white-label and OEM device options, or commit to new hardware because the wearable form is the product.

The repeatable sentence is simple: choose the wearable by mouth, mount, moment and mass, then choose the manufacturing path by how much of that shape must be new.