Five steps, one system: survey-grade records of every plant on the land, and a robotics stack built to make restoration affordable at any acreage.
Every step reads from the ledger and writes back to it. That's what makes step five cheaper than step three, and next year cheaper than this one.
Drone flights capture survey-grade RTK/PPK aerial imagery of the whole site. Below the tree canopy, where aerial imagery can't see, quadruped ground robots and field crews walk the understory to complete the map.
Every plant, native and invasive, gets a permanent ledger entry: position, species, condition. Our proof-of-concept terrain is majority oak woodland, so sub-canopy sensing is a requirement, not a gimmick.
From the catalog, our software generates the clearing plan, the mulch zones, and a native planting palette matched to the site's ecology: oak woodland, chaparral, or grassland. Everything is reviewable on a map before any work starts, plant by plant and zone by zone.
Ground crews remove invasive annual grasses and sheet-mulch at scale, guided plant-by-plant by the survey map. It is targeted work, not blanket scraping, and no burning. Crews handle the sensitive zones around existing natives; robotic mowing platforms take the open ground.
Natives go in the ground per the plan. Every planting is RTK-tagged into the ledger at the moment it happens: species, exact position, date. From that point forward the plant has a permanent identity, and every later observation appends to its history.
Seasonal drone flights re-survey the site. Because every intended plant is already mapped, anything unmapped is flagged as likely weed growth. The map makes detection sharper every season. Flagged weeds get targeted removal by crews and autonomous ground robots. Aerial micro-treatment sits on the research roadmap.
Plant health is tracked in the ledger over time, and maintenance effort declines as the natives close in.
Observations are never overwritten, only added. The full history of every plant is preserved: what was there, what we removed, what we planted, and how it's doing each season.
Each step above writes rows into the same record. The survey writes what exists. Clearing writes what was removed. Planting writes what went in. Monitoring writes how it's doing, season after season, against the same identities.
Because rows are immutable and current state is derived from history, the record can't be quietly rewritten. That turns restoration into auditable evidence: for compliance reporting, and for documented, maintained mitigation that holds up years later.
| PLT-00412 | Quercus douglasii | 37.4821, −119.9663 | PLANTED | 2025-11-02 |
| INV-09117 | Avena fatua | 37.4823, −119.9660 | REMOVED · INVASIVE | 2025-10-18 |
| PLT-00412 | Quercus douglasii | 37.4821, −119.9663 | OBSERVED HEALTHY | 2026-04-12 |
| PLT-00488 | Stipa pulchra | 37.4819, −119.9665 | OBSERVED HEALTHY | 2026-04-12 |
| PLT-00412 | Quercus douglasii | 37.4821, −119.9663 | OBSERVED HEALTHY | 2026-07-30 |
We label capability honestly because our buyers audit claims for a living. Here is the whole picture in one place.
Expanding autonomous treatment coverage until cost per acre supports thousand-acre restorations.
Because our proof-of-concept terrain is majority oak woodland, and the plants that matter most live where aerial imagery can't see: young oaks, understory natives, and the invasives crowding them. Legs handle slope, duff, and downed branches that stop wheels. The platform's job is simple: walk the understory, close the gaps in the map.
The process starts with a survey, and the survey starts with a walk. Tell us where your land is and what it's covered in. We'll bring the map.