Skip to content
AI NEWS#Planet AI#ai.planet.com#Planet Labs

Planet AI: Ask a Question, and the Satellites Show How a Place Changed Over Ten Years

Chatbots have read the entire internet, but they have never seen the world. Planet has opened an app to everyone that searches ten years of satellite imagery in plain language, with no GIS knowledge required. Here is what it can do, where it still falls short, and how to try it on a location in Israel.

Yaakov Tzedek
Yaakov Tzedek · AI expert·6 min read
Planet's announcement page for the Planet AI app, plain-language search across ten years of satellite imagery
THE BOTTOM LINE
  • Planet has opened Planet AI to everyone: a map with a chat interface that searches ten years of satellite imagery in plain language, compares dates, generates time-lapses, and runs Deep Research. Free for now.
DID YOU KNOW?

According to Planet, a beta user said they would never have manually scrolled across an entire desert to find something. Without plain-language search, some of those searches simply would not have happened. That is the point of the product: it does not replace an analyst, it makes it possible to ask questions no one would have bothered to check manually.

Chatbots have read the entire internet. They have never seen the world. Planet wants to change that, and has opened an app to everyone that searches ten years of satellite imagery for answers.

Planet is a company that operates a fleet of hundreds of small observation satellites and photographs the Earth's landmass nearly every day. On November 5, 2026, it opened Planet AI to the public: a map with a chat window where you type a plain-language question and the system searches its image archive for the answer. No GIS software, no course, and for now no payment either.

In short:

  • What it is: A natural-language search across an archive of more than ten years of Planet satellite imagery.
  • What you get: Locations that match a description, before-and-after comparisons, time-lapses, and a Deep Research mode that runs agents for complex analysis.
  • You can add your own data: Images or CSV files, and you can set up scheduled reports.
  • Price: Free for now. Planet hints that premium features are coming.
  • How to try it: Sign up at ai.planet.com. Works on mobile too.
  • The caveat: This is not a live feed, and it is not proof. It is an engine that finds visual evidence for you to verify.

What can Planet AI do?

Three things, according to the official announcement: search, explore, and act.

The announcement post on Planet's blog from November 5, 2026: Introducing the Planet AI App: Search. Explore. Act.

Screenshot: Planet's official blog (planet.com/pulse), 5.11.2026

  1. Search without knowing where to look. You describe what you are looking for, and the AI scans vast areas to find matching locations. Behind the scenes, according to Planet, language models interpret the request and query a global vector map built from AI analysis of visual features in monthly mosaics.
  2. See what changed and when. The system goes back in time, identifies when something changed, and presents a side-by-side comparison or time-lapse. Every result is backed by an actual Planet image you can open and inspect yourself.
  3. Analyze in depth. In Deep Research mode the system runs agents for complex analysis. You can upload external files, track trends, and set up scheduled reports.

Results can be shared as an image or a time-lapse, from a computer or a phone.

Who actually finds this useful?

Anyone whose question starts with "where" or "what changed." Planet reports that hundreds of users tried an early version over the summer, and most used it at the first stage of their work: figuring out where to look in the first place.

The examples the company gives:

  • Nature conservation: Teams check for encroachment into protected areas or fire risk, and verify automated deforestation alerts before sending rangers into the field.
  • Regulation and environment: Enforcement teams locate industrial sites across an entire district by their proximity to waterways and rank them for inspection. Work that used to take days of manual scanning, according to Planet, shrinks to minutes.

And in Israel, beyond organizations: journalists verifying construction or damage after a fire, appraisers and developers who want to see how a neighborhood developed, farmers, local authorities, and anyone simply curious about what happened to the coastline near their home.

The caveat you need to understand before using it

Planet AI is not a live camera, and it does not pass judgment. The broad search relies on monthly mosaics, meaning one composite image per month. That is excellent for changes over years, and weak for the question "what is happening there right now."

More importantly: when the system flags something that looks like construction, mining, or flooding, that is a lead, not a conclusion. It does not know who did it, why, or whether it is legal. The real advantage here is that the answer is not just text the model composed. Behind it is an image with a date, and you can check it yourself.

How to try it on a location in Israel

  1. Sign up at ai.planet.com.
  2. Ask a focused question, for example: "How has the Dead Sea shoreline changed from 2016 to today?" or "Where has construction been added on the outskirts of my city over the past five years?"
  3. Request a comparison or time-lapse between two dates.
  4. Open the image itself and check its date before drawing any conclusions.
  5. Choose non-sensitive locations. Do not query or share images of military bases, security installations, or critical infrastructure.

If you already use Deep Research in ChatGPT or Gemini, the logic is the same: a well-defined question produces a good answer. The guide to AI agents covers that principle in the context of day-to-day work.

Also this week in leading tools

Atmee v2: the next chatbot looks you in the eye

From a single photo and a short voice clip, Atmee v2 generates a live video character that listens, nods, and responds with facial expressions. The startup Atmanity launched the model on November 4, 2026, alongside a second model called Causeur that decides how the character behaves while Atmee renders the video. The company claims the model is 50 times faster and 10 to 100 times cheaper than competitors, and runs in real time on a single GPU. These are the company's own claims, not yet independently verified, and access is currently by invitation only.

The direction is clear, and we already saw it with Meta: Muse got a voice and a face ten days ago. After text and voice, agents are moving into video calls. For private tutors, customer service, and coaches, this is an opportunity. For deepfakes in live conversation, it is a threat, and it is worth discussing in your organization right now. Anyone who wants an avatar today can start with the HeyGen guide.

OpenAI and ModRetro: a game you described in words, on a real cartridge

OpenAI has put the ModRetro Chromatic: Codex Edition on sale, a Game Boy-style handheld console, for $200. The console comes with a cartridge you can burn games onto. You describe a game in plain language, Codex (OpenAI's coding agent) builds it, you test it in an emulator, and then you burn it. The collaboration was announced in September at OpenAI's developer conference, and public sales began on November 9.

This is not a new model, it is a toy. But it shows something real: people without a programming background are now building a complete product through conversation. Anyone who wants to get started with Codex without the console can find everything in the OpenAI Codex guide.

Want to learn how to work with agents that deliver results?

Planet, Atmee, and Codex all point in the same direction: give an agent a task in plain language and it returns a result you can verify. The AI agents bundle teaches exactly that: how to give an agent a real task, set its boundaries, and check what it did, using tools that work right now.

Sources

Frequently asked questions

Was this article helpful?
Share:WhatsAppLinkedInX
Yaakov Tzedek
WRITTEN BY

Yaakov Tzedek

An Israeli AI lecturer, consultant and implementer who has supported 100+ organizations in AI implementation and lectured for 200+. Official content partner of Meta, Google, Canva and TikTok, featured on Israel's Channel 13.

More about Yaakov →