Skip to content

Tree inspection software for UK arborists

Tree inspections, from field to finished report.

Capture tree data on site, even without signal. Review your findings, then issue a professional arboricultural report without rebuilding your notes at the end of the day.

  • Survey on site without signal
  • Nothing is issued until you approve it
  • BS 3998-oriented workflow
Tree record for T1, English oak: dimensions and location, base, stem and canopy photographs, description, condition, a table of recorded defects with severity, and recommended works with their basis.
A recreated app screen and a page of a real ARB-FLO report, both from the fictional sample survey.
  1. 1

    Field

    Trees, photos, GPS and voice notes, captured on site with or without signal.

  2. 2

    Structure

    Findings held as structured survey data, with rule-based candidate actions.

  3. 3

    Review

    You check every value, edit the draft wording and approve it.

  4. 4

    Report

    A numbered, professional PDF, ready to issue.

The problem

Less time turning field notes into paperwork.

Inspecting the tree is rarely the hard part. The hard part is everything around the inspection: keeping notes, photographs, locations and recordings with the right tree, then writing it all up.

  • Field notes

    Scattered notes, measurements and observations.

  • Photos

    Hundreds of images that need to stay with the correct tree.

  • GPS

    Locations that need to stay attached to the survey record.

  • Voice notes

    Useful observations captured on site, but hard to turn into structured report language.

  • Admin

    Hours spent converting field information into a professional report.

Back at the desk

  • IMG_4021.jpg
  • Voice 031.m4a
  • site visit notes
  • IMG_4022.jpg
  • map screenshot.png
  • T4 ??
  • IMG_4025.jpg
  • Report_v3_FINAL.docx
With ARB-FLO: one record per tree

T4

Fagus sylvatica (beech)

In review
Dimensions
1050 mm24 m20 m
Location
52.204905, 0.121730±7 m
Base photograph of tree T4 showing a cavity
Base
Stem photograph of tree T4
Stem
Canopy photograph of tree T4
Canopy
00:31
Cavities · ModerateDeadwood · Minor

How it works

From first tree to issued report, in one app.

Follow one survey through the five steps the inspector takes: three on site, two back in signal. The screens are recreated from the ARB-FLO app, using the fictional project behind the sample report.

Step 1 of 5

Create your project

Set the site up once: name, client, location, date, weather and the inspecting arborist. These details head every page of the report, so they are captured before the first tree.

  • Client, site, date and weather recorded once for the whole survey
  • Inspector name and qualifications carried into the report
  • Saved on the device if you are offline, then synced

Works without signal; syncs when you are back online

Voice-first field capture

Keep your eyes on the tree.

Typing into a phone while you stand under a canopy is slow and awkward. Say what you see instead. ARB-FLO proposes the structured values, each one quoted from your words, and you decide which stand.

  1. Voice
  2. Transcript
  3. Extraction
  4. You confirm
  5. Tree record

What was said

T4, over-mature beech, approximately 1050 millimetres DBH, about 24 metres tall, crown spread around 20 metres. Moderate cavity on the south side of the stem at about one and a half metres, and minor deadwood scattered through the crown. There might be some decay at the base. No obvious fungal brackets. I’d recommend a crown clean.

  • Becomes a value
  • Needs confirmation
  • Stated as not observed
  • Works mentioned

What ARB-FLO proposes

  • Tree ID

    T4

    From speech
  • Species

    Beech

    From speech
  • Age class

    Over-mature

    From speech
  • DBH

    1050 mm

    ≈ Approximate
  • Height

    24 m

    ≈ Approximate
  • Crown spread

    20 m

    ≈ Approximate
  • Cavities

    Moderate

    From speech
  • Deadwood

    Minor

    From speech
  • Basal decay

    Not graded

    Needs confirmation

    “There might be some decay at the base” Waits for your decision.

Stated as not observed

“No obvious fungal brackets”. Not recorded as a finding.

Works you mentioned

“I’d recommend a crown clean”. Not added to the survey.

  • Quoted, not invented

    Every value has to be supported by words in the transcript. Anything the transcript does not support is discarded.

  • Hedges are flagged

    “Might be” and “maybe” become Needs confirmation, not a recorded finding, until you decide.

  • Speech never becomes advice

    Works you mention are kept for reference. Recommended works come from knowledge-base rules and your own choices.

Live transcription uses your browser’s speech service where it is available and is switched on per device; otherwise you type what you said. Extraction understands English and common UK species names, and anything it cannot place stays in your notes for you to review. The capture screen states where audio and text are processed.

Offline-first

Built for the field. Even without signal.

ARB-FLO is a progressive web app designed around unreliable field connectivity. Survey data is saved on the device first and synchronised when connectivity returns.

  1. Online01

    Survey as normal

    Trees are saved straight away and appear in the inventory.

  2. Offline02

    Keep working

    No signal in the woodland? Keep surveying. Work is saved on the device and queued.

  3. Online03

    Sync

    Back in signal, the queue is sent in order and the inventory catches up.

Works with no signal

  • Create and edit projects and trees
  • Record GPS position, dimensions and defects
  • Add photographs
  • Record voice notes
  • Open projects and trees already loaded on the device

Needs a connection

  • Turning a recording into tree data
  • Drafting the report wording
  • Approving a tree
  • Generating a report
  • Reopening a tree created offline, until it has synced

Each change is queued with its own identifier and replayed in order, so a retry never creates a duplicate tree. A sync screen lists what is waiting and lets you retry or discard a change the server rejected.

Professional reports

From field data to a report you can actually issue.

The report is generated from the approved survey record, not retyped. Browse the pages of a sample report below, or download the whole PDF.

Report cover for the Riverside Park Tree Survey: a deep green band with concentric arcs, then the client, site, inspection date, inspector, report reference ARB-4F1C2A-01 and Issue 1.

Page 1 of 13. Client, site, inspector and a numbered issue on the cover. Select the page to open it full size.

What goes into a report

  • Project information. Client, site, date, weather and inspector, captured once and carried through.

  • Tree schedule. Every tree in one table, with its defects, recommended works and timescale.

  • Findings. Recorded defects with their severity, tree by tree.

  • Recommendations. Approved works, each with its basis: a knowledge-base rule or the inspector.

  • Photographs. Base, stem and canopy photographs alongside each tree record.

  • Location. A site plan drawn from the recorded GPS positions, with each tree’s accuracy stated.

  • Numbered issues. Each issue is numbered and stored with a snapshot of the data behind it.

  • Professional formatting. A consistent cover, contents, headers, footers and page numbers, as a PDF.

Issue 1
Numbered issue
13 pages
Cover to declaration
6 trees
Each with a record
Download the sample reportPDF · 0.5 MB · 13 pages

A real report generated by ARB-FLO from the fictional survey shown on this page: the client, site and inspector are invented, and the photographs are sample images. The inspector remains responsible for the content of every report they issue.

Want to see one built? A demo runs a survey from the first tree to the issued PDF.

Request a demo

Structured and deterministic

Facts, rules and wording, kept apart.

ARB-FLO holds three things separately: what you recorded, what the knowledge base says follows from it, and how it is written up. Rules decide the second. AI, where it is used, only drafts the third.

1 · yours

Facts

T4 · Beech (Fagus sylvatica)

Age class
Over-mature
DBH
1050 mm
Height
24 m
Crown spread
20 m
Cavities · ModerateDeadwood · Minor

Structured fields from the knowledge-base defect library. Only you change them.

2 · knowledge base

Rules

Knowledge-base rule: Deadwood

Why: Deadwood recorded

Crown clean / removal of deadwood

Cavities

The knowledge base defines no action for this. ARB-FLO says so and leaves the action to you, if you want one.

Fixed rules, so the same facts give the same suggestions. They are candidates: you include, edit or drop each one.

3 · AI-assisted, optional

Wording

AI draft

Unreviewed. Written only from the recorded facts.

T4 is an over-mature beech (Fagus sylvatica) with a DBH of 1050 mm, a height of 24 m and a crown spread of 20 m.

A moderate cavity was recorded on the south side of the stem at approximately 1.5 m. Minor deadwood was scattered through the crown.

Wording only. It never adds findings and never proposes actions, and you approve every word.

Mapped today

Restraint over invention

The rules use the knowledge base’s own wording, and ARB-FLO does not present anything as derived from BS 3998 unless the knowledge base says so. A recorded defect with no rule is never given an invented action: it is flagged, and the decision stays with the inspector.

  • DeadwoodCrown clean / removal of deadwood
  • Proximity to targetCrown reduction or lateral pruning
  • Severe decaySectional dismantle / Fell to ground level

Human-in-the-loop AI

AI assists. The arborist decides.

ARB-FLO does not give a language model control of the inspection. It keeps facts, recommendations and prose separate, and puts your approval between the draft and the report.

AI draft

Unreviewed. Written only from the recorded facts. Nothing here reaches a report until you approve it.

Condition

A moderate cavity was recorded on the south side of the stem at approximately 1.5 m. Minor deadwood was scattered through the crown.

Labelled as a draft, kept separately, and never copied into a report on its own.

Approved content

Inspector-approved final content

Approved by Alex Morgan on 3 October 2026, 14:20 UTC · drafted with AI assistance and reviewed by the inspector.

Recommended works

  • Crown clean / removal of deadwood · Within 3 months
  • Assess the extent of decay at the stem cavity · Within 3 months

Only content in this state is issued in a final report.

  • The model only sees the facts

    It is given the tree’s recorded facts and field notes, and nothing else. Client name, site address and coordinates are never shared with the AI provider.

  • Output is checked, not trusted

    An incomplete survey cannot be drafted or approved. Drafts must pass a strict schema, and checks flag figures, defects, decay, hazard or stability wording and recommendations the record does not support.

  • It never proposes actions

    Recommended works come from knowledge-base rules and your own choices. The model does not add findings, grade severity or assess risk.

  • The report says so

    Where AI-assisted wording is used, the report’s limitations state it, and that the inspector reviewed and approved it.

AI drafting is optional. With no AI provider configured, ARB-FLO assembles the draft from your notes with a template and labels it as a template, never as AI-assisted.

Reduce admin. Don’t replace professional judgement.

Features

Built around field arboriculture, not a generic workflow.

Fifteen capabilities, each one something the app does today.

01

Capture in the field

Everything you gather on site, kept with the right tree.

  • Offline-first surveys

    Create projects and trees, add photographs and record voice notes with no signal. Changes queue on the device and sync when you are back online.

  • GPS location capture

    Take a GPS fix with its accuracy shown, or type coordinates in. The position stays with the tree.

  • Tree inventory

    Every tree in the project in one list, with species, stem diameter, height and status at a glance.

  • Voice notes

    Record a note against each tree, play it back, replace it or delete it. Recordings never appear in the report.

  • Photo capture

    Base, stem and canopy photographs for each tree, orientation-corrected and prepared for the report.

02

Structure and decide

Facts, rules and wording kept apart, with you in charge.

  • Structured tree data

    Species, age class, dimensions, defects, target and notes held as fields, so every survey has the same shape.

  • Knowledge-base-driven findings

    Defects come from the knowledge-base library rather than free text, so a finding is recorded the same way every time.

  • Deterministic recommendations

    Candidate actions follow fixed knowledge-base rules and show their reasoning. Where there is no rule, none is invented.

  • AI-assisted prose

    Drafts report wording from the recorded facts only. Output is checked against the record and labelled as a draft. It never proposes actions.

  • Human review

    Nothing reaches a final report until you have reviewed, edited and approved it.

03

Issue and integrate

A report you can stand behind, and data that is not locked in.

  • Professional PDF reports

    Numbered reports with a cover, contents, site plan, tree schedule, tree records with photographs, limitations and declaration.

  • Audit trail

    Issued reports are immutable and numbered, stored with a snapshot of their data and a SHA-256 checksum. Drafts are kept apart from approved text.

  • Data export

    Reports download as PDF. The survey data behind them is structured and available through a documented REST API.

  • Progressive web app

    Install from the browser to your home screen. No app-store account, download or review queue.

  • Integrations and webhooks

    Signed events fire when a tree is approved and when a report is generated, so other systems can follow along. An example n8n workflow is included.

Who it’s for

For people who inspect trees and answer for the report.

From the sole practitioner to a practice with systems of its own. Designed around UK inspection practice: report wording uses the terminology of BS 3998:2010, in British English and metric units.

  • Independent arborist

    Spend less time writing up surveys.

    • Survey, review and report from one phone
    • Your approval, your declaration, your numbered report
  • Arboricultural consultancy

    Standardise field data and reporting, survey after survey.

    • One defect vocabulary and one report structure
    • Every issue numbered the same way
  • Tree surveyor

    Capture structured information without slowing down in the field.

    • Voice, photographs and GPS kept with the right tree
    • Works on site with no signal
  • Larger practice

    Build a consistent digital workflow around your existing systems.

    • Signed events when trees are approved and reports generated
    • A documented REST API for the survey data

ARB-FLO is built around a single inspector’s workflow today. Sign-in and shared team accounts are not part of the product yet, and we will say so plainly when we scope a pilot with you.

Demo and pilot programme

Spend less time writing up tree surveys.

See ARB-FLO run through a survey, from field capture to the issued report, or ask to join the pilot programme.

What happens next

  1. 1

    You send the form. A name and an email address are all we need. No account, no payment details.

  2. 2

    We reply by email. To arrange a time that suits you. We use your details only to reply.

  3. 3

    We walk through a survey. From the first tree to the issued report, and talk honestly about whether ARB-FLO, or the pilot, fits how you work.

Request a demo

Or ask about the pilot, or send us a question.

I’m interested in

Required. We use these details only to reply to your request; the form sends them to us through Netlify, which hosts this website. See our privacy policy (opens in a new tab) for what happens to them.