Live imagery chat demo

Talk to satellite, aerial, and drone imagery in plain English.

naipai shows what happens when imagery becomes a conversation layer: ask what changed, count assets, find damage, summarize a site, and get answers with confidence and caveats.

This page is a demo. The embedded version uses sample aerial imagery, but the same workflow can be customized for any image source, inspection process, output format, and buyer use case.

Try the Demo Scope a Custom Version

Live sample demoBuilt around NAIP-style aerial imageryCustom builds can use your imagery and workflows

Imagery conversation preview

Sample scene

User question

Which parts of this site need a field check first?

naipai answer

Priority review should start along the eastern edge, where several features appear partially obscured and access conditions are less clear.

Confidence: Medium. Caveat: confirm with a second date or field photo before dispatch.

Find visible conditions

Export next actions

Live sample

Try the demo, then imagine it on your imagery.

The embedded app is intentionally a sample. It proves the core interaction: a user can ask imagery a question and get a usable answer. A production version can use your sensor, your archive, your field workflow, your review language, and your delivery format.

Open demo in a new tab

Why it matters

This is more than a chatbot. It is an interface for visual work.

The value is not the conversation itself. The value is making hard-to-review imagery searchable, explainable, repeatable, and useful for decisions your team already has to make.

Ask the image directly

Replace manual panning, zooming, and note-taking with plain-English questions about what is visible in the scene.

Turn visual review into outputs

Counts, condition notes, change summaries, severity rankings, QA checklists, report sections, CSVs, and map-ready annotations.

Keep confidence visible

Every useful answer should tell the user what was found, how confident the system is, and where a human should review.

Customizable by design

Bring any imagery. We shape the assistant around the job.

If your team can define what a good visual review looks like, we can usually adapt the workflow around it. The demo is a starting point, not the limit of the product.

Any imagery source

NAIP, Sentinel, Landsat, commercial satellite, drone orthomosaics, inspection photos, thermal, multispectral, scans, or customer-owned archives.

Your questions and rules

We tune the prompt, schema, examples, confidence language, thresholds, and review process around the decisions your team already makes.

Your output format

Answers can become a map layer, report, ticket, CSV, GeoJSON, dashboard card, API response, or analyst review queue.

Private deployment path

Custom versions can include private data, authentication, role-based access, audit trails, human review, and integrations.

Practical use cases

The strongest uses are repetitive, visual, and expensive to inspect manually.

Damage and response

Find likely roof damage, flooding, debris, blocked access, exposed assets, and priority review zones after storms or incidents.

Which structures show visible damage?

Rank the highest-risk parcels in this scene.

What changed compared with the earlier image?

Asset inventory

Count and classify visible assets such as solar panels, tanks, equipment yards, pools, rooftop systems, vehicles, and utility features.

Count solar panels by roof section.

Identify equipment staging areas.

Flag assets that need a field check.

Land and vegetation monitoring

Screen fields, corridors, parcels, and right-of-way areas for canopy, stress, encroachment, disturbed ground, or land-use changes.

Where is vegetation encroaching?

Summarize land-use mix in this image.

Which areas look newly disturbed?

Inspections and compliance

Convert repetitive visual inspections into repeatable review flows with evidence, limitations, and exportable notes.

Does this site match the expected condition?

List visible compliance concerns.

Create a field follow-up checklist.

Pilot path

A custom version starts with your highest-value visual question.

We are not limited to satellite imagery. The right pilot starts with the image type and decision workflow that matters to your customer, operation, or internal team.

Bring real imagery

Send a small sample set from the actual work: drone captures, satellite scenes, inspection photos, or historical imagery.

Pick valuable questions

Choose the 3-5 questions that save analyst time, reduce field visits, create a report, or support a customer-facing product.

Build the workflow

We configure the model behavior, answer structure, confidence cues, evidence panels, exports, and user path around your use case.

Ship a working pilot

The result can be a private demo, embedded app, API endpoint, analyst dashboard, or client-ready report workflow.

What a real implementation can include

A private imagery assistant your team can actually use.

The production opportunity is bigger than this public demo: custom image ingestion, geospatial context, answer schemas, audit-ready outputs, review queues, and integrations with the systems where work gets assigned.

Review time

hours to minutes

Sources

sensor-agnostic

Outputs

report/API/map layer

Users

analyst or customer

More working examples

Explore More of Our Demos

Compare this sightline scoring tool with the full collection of mapping and geospatial decision tools we've built.

Browse All Demos

Want naipai on your imagery?

Bring a sample image set and the question your users keep asking. We will scope a custom imagery-chat workflow around your data, your confidence requirements, and the output your team needs.

Book a Custom Imagery Briefing See More Web Map Demos

Implementation stack

Technologies We Work With

We choose the stack around the job: browser maps for decision surfaces, spatial databases for reliable data, Python and GIS libraries for analysis, and annotation tools when imagery needs reviewer-ready labels.

QGIS

GIS Software

ESRI ArcGIS

GIS Platform

PostgreSQL

Database

PostGIS

Spatial Database

AWS

Cloud Platform

Google Cloud

Cloud Platform

DuckDB

Analytics Database

OpenAI

AI Platform

Claude AI

AI Assistant

CVAT

Annotation Tool

Python

Programming

React

Frontend

Node.js

Backend

Docker

Containerization

Kubernetes

Orchestration

Azure

Cloud Platform

TensorFlow

Machine Learning

Pandas

Data Analysis

NumPy

Scientific Computing

Jupyter

Data Science

Git

Version Control

Linux

Operating System

Ubuntu

Operating System

Mapbox

Mapping Platform

Leaflet

Web Mapping

Fastapi

API Framework

GeoPandas

Geospatial Analysis

GDAL

Geospatial Library