Akaro: Soil to Harvest

Topsoil: the app in action

See it work.

Point, shoot, plant. Watch Akaro read three different soils.

Fix it at home

Akaro suggests natural fixes you can make yourself, like compost, rice husk ash, aged manure or biochar, so you can grow what you actually want.

Grow what fits

It also ranks the crops that already suit your soil today, so you can start planting right away.

Sharper with your location

Switch on location and Akaro can refine each reading with local rainfall, elevation and regional soil patterns.

Planned for the app. Today your location already shapes the weather forecast.

Illustrative demo with example values, not real readings.

9:41

Scanning soil

Proof

From a handful of dirt to a plan.

Drag the handle. Same soil, same photo. One side is a guess, the other is a report.

Before: a guess

Color: dark brown Texture: crumbly loam Moisture: good After: a report

Sample soil report

Dark loam

pH6.2

Organic matter3.8%

DrainageGood

Top crop matches

Chili92%

Rice81%

Nutmeg74%

Next step
Mix in 2 kg of compost per square metre before planting.

Illustrative example, not a real measurement.

Topsoil: how I created Akaro

From a photo to a planting plan.

Akaro is a real web app with real sign-up, not a mock-up. This is the path every photo takes.

A photo goes in

You sign in and upload a picture of your soil. It's saved to your own account, and the database only lets you see your own data. You can delete it any time.

AI reads the soil

A vision model studies color, texture, moisture and visible organic matter, then returns a structured report: estimated pH range, drainage, fertility, and 3 to 5 natural remedies like compost, rice husk ash, aged manure or biochar.

Plants are matched to the soil

The report is checked against a growing database of crops, each with soil, climate, water needs and days to harvest, in both languages.

Weather shapes the advice

A live five-day forecast for your location becomes plain gardening guidance, and your garden and growth log keep it all in one place.

Next.js 15SupabasePostgres + row-level securityGemini + Groq AIOpen-MeteoTailwind CSS

Topsoil: the app itself

The app, up close.

Six screens, one job: help you plant with confidence.

Photograph your soil

Point your phone at a handful of earth. Akaro scans it and gets to work.

A reading, then a plan

Texture, moisture and color, then plants ranked for your soil, each with a reason.

A library that keeps growing

Search more than 13,000 plants with their soil, climate and water needs.

Keep a garden log

Track what you plant and note what you do, so next season starts with notes, not memory.

Weather that guides

A local forecast turned into plain advice for your plot.

In your own language

Switch between English and Bahasa Indonesia with one tap.

Screens show sample data.

Everything it does, in one place.

Built for a phone in a field: quick, clear, and in your language.

Soil scan

One photo in. A full reading out.

Akaro looks at color and texture, then estimates pH, organic matter and nutrient levels, with a plain-language fix for each.

Organic

Nitrogen

Moisture

Plant library

13,000+ plants, matched to your soil.

Search the library or let Akaro shortlist what fits.

RiceChiliCassavaGingerTomato

Weather

Advice that watches the sky.

Forecast-aware tips, like holding water before heavy rain.

MonSun

TueRain

WedRain

Garden log

Watch your garden grow.

Keep notes and photos for every plot, season after season.

Two languages

English and Bahasa Indonesia.

Switch any time, everywhere in the app.

ENID

Two AIs, one answer

Gemini and Groq, working together.

Gemini reads the photo and Groq's fast models help match plants, so answers come back quickly.

Always at hand

Installs to your home screen.

Open it like any other app, and a friendly offline page keeps it from breaking when the signal drops.

Topsoil: what it's for

Watch a plot thrive.

Same seed, same twelve weeks. One plot is guessed at. The other follows Akaro's advice.

Without guidance

With Akaro

Week 0

Week 1The photo scan reads acidic, low-organic soil
Week 3Akaro suggests compost and dolomite lime
Week 6Forecast guidance: hold the water before heavy rain
Week 9The growth log shows steady gains
Week 12Harvest, and the data feeds the next season

Illustrative growth to show the idea, not measured data.

Topsoil: why it matters

A soil check that costs nothing.

Testing soil in a lab is the gold standard, but for most small farms it means digging samples, getting them to a lab, paying, and waiting. Akaro gives a free first look in the field.

A routine lab check, per sampleRp 414,000

Akaro, per photo

Rp 0

Dig
Travel
Pay
Wait
Decode

One photo

5 steps become 1Dig, travel, pay, wait, decode. Akaro replaces all five with one photo.

No trip, no shippingThe answer comes to the field instead of the sample going to a lab.

Plain languageResults in English or Bahasa Indonesia, with a planting plan, not a table of numbers.

Samples you'd test per season: 5

Rp 2,070,000

saved per season compared with lab fees

Lab rate: Rp 414,000 per sample for a routine soil chemistry analysis, as billed in the 2023 report of the BPSIP South Sulawesi government lab (Ministry of Agriculture). Prices vary by lab. The leanest option, preparation plus pH, organic carbon and nitrogen, is Rp 96,000 at BSIP Yogyakarta. Shipping and travel are not included, so real costs can be higher. Akaro's photo analysis is an AI estimate and a quick first look. A lab measures things a photo cannot, such as available phosphorus and potassium, so for big decisions a lab test is still worth it.

Subsoil: the data underneath

Data science, from photo to plan.

Akaro turns a messy photo into clean, comparable data, joins it with other datasets, and ranks what to plant.

Under the hood

What runs behind every scan

13,000+

plants in the Akaro plant library today

Including 1,100+ researched with full soil, climate and water details, and growing every day

1,0005,00010,00013,000+ today

Growing every day. Progress shown on a log scale.

Photo to structured data

A vision AI model reads each soil photo and returns the same fixed fields every time, so every analysis can be compared and tracked.

{ "soil_color", "texture",
  "estimated_ph_range",
  "organic_matter", "drainage",
  "fertility_notes",
  "natural_remedies": [ ... ],
  "confidence_note" }

A plant knowledge base

Assembled from public plant datasets such as GBIF and USDA plus researched entries, stored in two languages, and extended automatically when someone searches for a species it doesn't have yet.

Matching and scoring

The soil report and the farmer's location are scored against the plants in the library that have full growing details. The six best crops come back, each with a one-line reason.

Live weather

A five-day forecast from Open-Meteo is joined in and turned into plain-language guidance for your garden.

Privacy by design

Row Level Security in the database means each person can only read and write their own analyses, garden and growth log. The AI key never reaches the browser.

Two languages, one dataset

English and Bahasa Indonesia live side by side in the same records, so advice reads naturally in both.

1. Photo

Unstructured pixels

2. Features

Color, texture, pH, organic matter, drainage

3. Join

Crop dataset and weather forecast

4. Score

Rank every crop by fit

5. Advice

A plain-language plan

Live model: build your own soil

Drag the sliders. Every crop is re-scored instantly with a simplified, illustrative scoring model, and the math is shown below.

Soil pH Organic matter Drainage

Decision map

Each cell shows the best crop for that pH and drainage, at your organic matter level. Brighter means a better fit. The white ring is your soil.

Try it: pick a soil

Best crops for this soil

fit = 0.5 x pH match + 0.3 x drainage match + 0.2 x organic matter match

A simplified, illustrative model with example crop needs. Amber names are native Indonesian plants. The app compares the photo report with its full crop database.

Feature space: where your soil lands

Each bar is a crop's preferred pH range, placed at its preferred organic matter. The closer your soil lands to a bar, the brighter it glows. Amber bars are native plants.

Pictures become features

Every analysis returns the same fields, so soils can be compared and tracked over time.

Datasets get joined

Soil readings, crop data and live weather work together instead of sitting apart.

The dataset keeps growing

A plant that isn't in the database gets researched and saved, so real use makes the data richer.

Subsoil: what belongs here

Rooted in Indonesia.

Most plant tools stop at the crops everyone already grows. Akaro also focuses on Indonesia's indigenous and traditional plants, the ones adapted to local soil and rainfall, so they get a place in the data too.

Pala

Nutmeg

Native to the Banda Islands in Maluku, and a spice that shaped world history.

Cengkih

Clove

Native to the Maluku islands, where clove trees grew long before they were traded.

Sagu

Sago palm

A staple of Papua and Maluku that thrives in wet, swampy soil where other crops struggle.

Salak

Snake fruit

A palm fruit native to Java and Sumatra, with scales like a snake's skin.

Temulawak

Javanese turmeric

A rhizome from Java, used for generations in traditional jamu.

Manggis

Mangosteen

A fruit tree of the Malay archipelago, loved across the region.

Every plant in Akaro carries its Indonesian name, and we're growing the database to include many more native species, so recommendations reflect what actually belongs in this soil and climate.