Silva PerpetuaGuides

Import anything

Bring in a spreadsheet, CSV, Word or PDF file, a photo of a field sheet or a map file. The app proposes how to read it, converts every value itself, shows you the result beside the original, and stores nothing until you confirm.

For
Consultant foresters
Time
8 minutes
Where
Web app — Stand › Field work › Imports, or the property page
Updated
On this page
How it works
  • The AI reads the file and proposes a mapping — how to read it — never the data itself. The app’s own engine then converts every value and shows the result next to the original cell.
  • Nothing is guessed: units, girth vs diameter, date formats and coordinate systems become questions with pickers.
  • Nothing is stored until a person confirms. Imported values keep the import as their origin.
  • Save the mapping as a format, and the next file with the same headers is read without AI.

Upload a file

Web appField work › Imports › Import anything (also on the property page).
  1. 1 Choose the file
  2. 2 Saved format, or match by headers
  3. 3 Ask Claude to propose the mapping
  4. 4 Read the file
  1. Choose the file — spreadsheets (every sheet), CSV, JSON, Word, PDF, a photo of a field sheet, or a map file (GeoJSON, zipped Shapefile, KML/KMZ, GPX, GeoPackage).
  2. Pick a saved format, or leave Match by headers.
  3. Tick Ask Claude to propose the mapping (untick it to write the mapping yourself or use a format).
  4. Choose Read the file. The import opens below, as a draft.

Review the preview

Web appA tally sheet in review.
  1. 1 Nothing is stored until you confirm
  2. 2 What confirming would store
  3. 3 Questions — nothing is guessed
  4. 4 Original → converted, row by row
  1. Nothing is stored yet — the notice says what to do next.
  2. Confirming would store — records, visits, new plots — and how many rows have errors. What changed lists the differences from the previous mapping version.
  3. Answer the questions. Here: “Larch” could be European, Japanese or hybrid larch — which is it? Each answer is a button; the preview updates.
  4. Read the table — each sheet, how each column is read (Girth (cm): DBH: GIRTH in cm → ÷ π → diameter, stored in m), and every row as original → converted, with its problems.
Web appAsk for changes, then confirm and save the format.
  1. 5 Ask for changes in plain words
  2. 6 Confirm — import these values
  3. 7 Save the mapping as a format
  1. Ask for changes in plain words — “Plot 8 was measured with a girth tape”, “skip the Notes column”. A new mapping version is proposed and the preview recomputed.
  2. Choose Confirm — import these values. Tick to create the new sample locations it found, or to leave out rows with errors. Confirming records you as the reviewer. Discard throws the import away.
  3. Save the mapping as a format with a name you will recognise.

Data that fits no method becomes a dataset

Cull records, habitat maps, bird and flora records fit no protocol. They become a dataset: a map layer with a style picker, a filterable table, and per-year totals beside the time series — context, never a score.

Web appA confirmed import: 26 rows became a dataset.

See Map layers and datasets.

Standard imports

Web appField work › Imports › Standard imports.
  1. 1 Format: tree tally CSV, GeoJSON, derived values
  2. 2 Diameter or girth, and the unit — never guessed
  3. 3 Upload for review (held as pending)
  1. Choose the format — tree tally CSV, GeoJSON, derived values from LiDAR, drone or satellite.
  2. Say which column holds the diameter or girth and in what unit — the app shows how it will read it (“DBH” read as DIAMETER in cm — no conversion).
  3. Choose Upload for review. The upload is held as pending until someone confirms it.
File content is sent to the AI provider

When Ask Claude is ticked, the file’s content is sent to Anthropic to propose the mapping. Untick it for files that must not leave your organisation, and use a saved format or write the mapping. See Your data.

What it doesn’t do (yet)
  • If AI import is not configured on your site, the chat says so: answer the questions, pick a saved format or write the mapping — deterministic imports still work.
  • There is a daily limit on AI-assisted imports per organisation; it resets each day.
  • No tree counts from imagery (planned: proposed values from supplied orthophotos).
Switch from spreadsheetsMove years of tally sheets and inventory spreadsheets into the app stand by stand, so your history becomes the first points of each stand’s trend — without retyping.6 minutes, then about 15 minutes per standMap layers and datasetsPut drone orthophotos, satellite or LiDAR tiles and public constraint layers on the property map, use them as exclusions for plot layouts and as landscape context, and explore datasets such as cull records.5 minutesReview what came backEverything assisted or imported — voice-note proposals, camera and instrument values, imports, a marking session’s intervention — waits for a person to accept, edit or reject it. Here is where to find it and how to decide.6 minutesYour data: export, transfer, privacy and AI processingThe client owns their woodland data. How to export it all, move a client to another organisation, what is stored where, and what is sent to AI services and when.5 minutes