The platform

One robot. Many crops. Built to adapt.

Most harvesting robots are designed around a single crop and a single farm. Ours is a modular, autonomous platform — one rover, interchangeable arms and grippers, and an AI mind that re-learns rather than being re-engineered. The result is a machine that moves between crops, seasons and sites instead of being replaced by them.

A self-contained field machine. Each unit pairs an autonomous rover with one or more robotic arms, edge-AI perception and the safety systems needed to work alongside people and plants. It navigates rows, finds ripe fruit, picks it gently, and keeps going — through changing light, weather and ground.

Because the intelligence lives on the machine and improves with use, the platform is not frozen at the moment it ships. It is autonomous, self-learning and self-improving by design: the same hardware gets better at the job the longer it does it.

What makes it one platform

  • Modular arms and grippers that swap to suit the crop in front of them
  • A common control and perception core shared across every configuration
  • Fleet coordination, so a group of robots works — and learns — as one
  • Gentle handling tuned for delicate, high-value soft fruit
Capabilities

The building blocks of an autonomous harvest.

Each capability is useful on its own. Together they make a machine that can take responsibility for a row, a polytunnel, a season.

Autonomous rover

Navigates rows and tunnels on its own, plans its route, and returns to charge without a driver. Built to keep working in the wet, the heat and the dust of a real growing site.

Self-driving · All-weather

Modular arms

A common arm design scales across crop weights and reach. Run one arm or several on a single rover, and coordinate them so they pick faster without getting in each other's way.

Multi-arm · Coordinated

Swappable grippers

A tool-changer lets the robot pick up the right end-effector for the crop — and put it down for the next. Gentle, fruit-specific handling keeps bruising low and grade high.

Tool-change · Gentle pick

Edge-AI perception

On-board 3D vision finds fruit, judges ripeness and plans the approach in real time — without waiting on the cloud. Seeing happens where the picking happens.

3D vision · On-device

Fleet coordination

Robots work as a group, sharing the field between them and a single supervisor. What one machine learns, the whole fleet can benefit from — the platform gets smarter at scale.

Swarm · Shared learning

Safety & traceability

Hardwired safety so the machine can share space with people, and per-pick records — what was picked, when, from which plant — that turn a harvest into clean, auditable data.

Safe-by-design · Per-pick data
Phase-one crops

Six crops to prove the multi-crop promise.

We are starting with a deliberately varied set of soft, high-value crops — different shapes, different handling, different plants. One platform that learns all six is a platform that can learn the next sixty.

Strawberry Cherry tomato Raspberry Runner bean Mini pepper Courgette

Why multi-crop matters

  • One machine earns across more of the year, not just one short window
  • Growers with mixed crops get one system instead of many
  • Every crop the platform masters widens the moat against single-crop rivals
  • Shared learning means each new crop starts further ahead than the last

See the platform in your rows.

We are selecting a small number of grower partners for the first field trials. Early access, real influence on the product.

Talk to us