# QAI

> QAI is the data and cloud layer under agentic AI products — one Postgres carrying records, documents, spatial indexes and vector columns, an agent runtime with retrieval and conversation memory, and per-product credit metering. It is built and operated by QAI Cloud, whose research arm is QAI Lab, and four QAI products run on it today.

Website: https://qai.io
Contact: info@qai.io
Legal name: QAI Cloud. Platform brand: QAI. Research arm: QAI Lab. Founded 2025.

## The platform

The layer is not sold as a standalone product today — it is the substrate under QAI's own products, and it is documented because what it does (and does not do) decides what those products can honestly claim.

- The QAI platform (https://qai.io/platform) — the data and cloud layer, in four named layers.
- Data (https://qai.io/platform#data) — Storage was never the problem. Disagreement is. An agent's knowledge is usually spread across systems that quietly disagree: what it retrieves sits in one, what the business considers true sits in another, what it remembers about a customer sits in a third, and who is allowed to see any of it sits in a fourth. An agent is only as right as the least fresh copy it can reach. So it reads the older one, and your customer is told something confidently wrong by a system in which every component passed its tests. Here, what is retrieved, what is recorded and what is remembered are one thing under one set of permissions — so there is no older copy to read.
- Agentic AI (https://qai.io/platform#agentic-ai) — You cannot review an agent's code and know what it will do. Same code, same permissions, a different prompt — and it takes a different action. Reading the source told you what a program would do; it tells you almost nothing about an agent. So the guarantee cannot come from the review it passed before it shipped. It has to come from what the agent is allowed to touch while it is running, and from what is kept about what it did. That is a different job from serving a request, which arrives and is gone, and a different job from running a deployment, which sits there until you change it.
- Cloud (https://qai.io/platform#cloud) — A budget an agent cannot exceed has to live below the agent, not inside its prompt. You are not shipping an app any more. You are handing a customer something that spends. The first day it runs on someone else's behalf, three questions arrive together: whose data was that, who pays for these tokens, and what stops it spending more. A cloud that bills you for machines cannot answer any of the three, so attribution becomes a spreadsheet that grows with every customer you win. Here the unit is the task — one agent's run, attributed to the customer it ran for, charged against a ceiling that was set outside it.
- Infra (https://qai.io/platform#infra) — Defaults are chosen at the bottom, or they are not chosen at all. Execution that is untrusted until proven otherwise. Outbound traffic that is closed until something opens it. A contained environment thrown away after a single task. None of these is a feature you install on top of primitives designed for a trusted person who deploys once a week — a default is what happens when nobody configures anything, so it gets decided underneath everything, before anything above it exists. This is also what makes running the whole thing again cheap, which is where the circuit starts over.

## Products

- QAI Finance (https://qai.io/finance) — AI-native research boards over public-market fundamentals: argued bull/base/bear theses with explicit assumptions, company intelligence, funds, and the AI value chain. Analyst work product.
- QAI Earth (https://qai.io/earth) — geospatial intelligence for climate and geoscience.
- QAI Insights (https://qai.io/insights) — market radars over the sectors AI is reshaping, with sourced traction and citations on the rows that carry a claim.
- QVC / QAI Ventures (https://qai.io/vc) — AI-native, creator-led venture program with rolling pitch intake.
- QAI Apps (https://qai.io/apps) — the catalog of vertical AI products QAI Lab builds on the QAI platform.

## Research & data

- Frontier AI research (https://qai.io/research)
- LLM leaderboard (https://qai.io/bench/rankings) and LLM API leaderboard (https://qai.io/bench/rankings/llm-apis)
- vLLM vs SGLang on one DGX Spark (https://qai.io/bench/edgebench/dgx-spark-qwen3.8-27b-vllm-vs-sglang): a recorded Qwen3.8-27B run — output and prompt speed, time to first token, and speculative-decoding acceptance, replayed token by token
- The AI stack, layer by layer (https://qai.io/insights/ai-stack) and the AI value chain as a network (https://qai.io/insights/ai-stack/value-chain)

---

# QAI Earth Documentation

> QAI Earth is a spatial intelligence platform for geoscience. Ask natural language questions about climate, geology, and Earth systems — with real-time data visualized on interactive maps.

URL: https://qai.io/earth
Documentation: https://qai.io/docs/geo

## What is QAI Earth?

QAI Earth is a geospatial intelligence platform that combines large language models with real-time Earth observation data. It enables anyone — researchers, analysts, students, or AI agents — to query geographic and environmental data using natural language.

Results are visualized as interactive map layers with customizable markers, and can be stacked to build comprehensive spatial analyses. The platform includes 8 built-in real-time monitoring layers, planetary trend charts from authoritative scientific sources, AI-generated reports, and collaboration tools.

## Core Features

### Natural Language Queries
Ask questions in plain English about any location, climate pattern, or geographic phenomenon. The AI understands spatial context and returns precise, map-ready data with source citations.

Example: "Show me active volcanoes near tectonic plate boundaries in Southeast Asia"

### Interactive Map Visualization
View results on a multi-theme interactive map. Switch between standard, satellite, terrain, and 3D Earth globe views. Dynamic markers with customizable colors, shapes, and sizes.

Map themes: Standard, Satellite, Terrain, 3D Earth Globe (WebGL)

### Data Layers
Every query creates a toggleable data layer. Stack multiple layers to build comprehensive spatial analyses. Customize marker styles and save your configurations.

Example: Combine deforestation, rainfall, and population layers for environmental impact analysis.

### Planetary Trend Charts
Track historical trends for key Earth metrics from authoritative sources including NASA, NOAA, and the World Glacier Monitoring Service. Data spans from 1980 to present.

Available datasets:
- Surface Temperature (°C Anomaly) — Source: NASA GISTEMP
- Atmospheric CO₂ (ppm) — Source: NOAA GML
- Methane Levels (ppb) — Source: NOAA GML
- Global Sea Level (mm Rise) — Source: NASA Sea Level
- Ocean Heat Content (10²² Joules) — Source: NOAA NCEI
- Glacial Mass Balance (m w.e.) — Source: WGMS
- Global Forest Cover (% Index) — Source: Global Forest Watch

### AI Report Generation
Generate comprehensive reports from your analysis sessions. Reports include executive summaries, key findings, location analysis, methodology, and recommendations. Export as PDF.

Report types: Business, Research, Environmental, Infrastructure, Resources, Tourism, General.

### Sharing & Collaboration
Share your analyses, data layers, and reports with public links or within your organization. Set view or edit permissions and optional expiration dates.

Share methods: Public link, Organization-wide, User-to-user.

## Chat & AI Queries

QAI Earth uses a conversational interface where every question creates a persistent chat session. The AI maintains context across messages, so you can refine your analysis with follow-up questions.

### How a query works

1. **Ask a question** — Type a natural language query about any location or Earth science topic.
2. **AI responds** — The response streams in real time with source citations from authoritative data.
3. **Locations extracted** — The AI automatically extracts locations, geocodes them, and places markers on the map.
4. **Data layer created** — A toggleable data layer is saved to your chat so you can revisit or combine it with other layers.

### Chat features

- **Multi-turn conversations**: The AI remembers your previous messages within a chat session (up to 20 messages of context). Ask follow-up questions to refine results without repeating yourself.
- **Message regeneration**: Regenerate any assistant message to get a fresh answer. Each version is saved so you can navigate between them.
- **Version history**: Every regeneration creates a new version with its own data layer. Navigate between versions to compare different analyses of the same question.
- **Source citations**: Responses include citations from Google Search grounding, linked to authoritative sources like NASA, NOAA, and scientific institutions.
- **Chat history**: All conversations are saved and searchable. Switch between chats from the sidebar, or start a new one at any time.
- **Suggested topics**: New chats display rotating topic suggestions across categories like mining, energy, climate, and technology.

### Reports from chat

After building up an analysis through conversation, you can generate a comprehensive report directly from your chat. The report synthesizes your queries, the AI responses, and the data layers into a structured document with:

- Executive Summary
- Key Findings
- Location Analysis
- Detailed Analysis
- Methodology
- Recommendations
- Conclusion
- Embedded Map Screenshot

Reports can be exported as PDF.

## Map Views

Switch between four map themes at any time from the settings panel or display mode selector.

- **Standard** (`map`): OpenStreetMap basemap with street-level detail. Best for urban analysis, infrastructure, and general navigation.
- **Satellite** (`satellite`): High-resolution satellite imagery. Ideal for land use, vegetation, terrain features, and remote area analysis.
- **Terrain** (`terrain`): Topographic map with elevation contours and hillshading. Best for geological analysis and elevation-dependent queries.
- **3D Earth** (`earth`): Interactive 3D globe rendered with WebGL. Supports rotation, zoom, and orbit controls. Best for global-scale visualization and satellite tracking.

## Data Layer Marker Customization

Every query result creates a data layer with location markers on the map. You can customize how these markers look from the Data Layers card.

- **Colors**: Pick from a palette of preset colors or choose a custom color for each data layer. Colors are saved per layer and persist across sessions.
- **Shapes**: Choose from 6 marker shapes — Circle, Square, Diamond, Triangle, Star, Hexagon — to visually separate different types of data.
- **Sizes**: Adjust marker size to emphasize important layers or reduce visual clutter. Icon sizes range from 8px to 48px, shape sizes from 20px to 80px. Preferences are saved to your account.
- **Layer controls**: Toggle individual layers on or off, rename, restyle, or delete them. When multiple layers are active, the map stacks them for side-by-side comparison.

## System Data Layers

Nine built-in monitoring layers provide continuously updated geospatial data from authoritative sources. Each layer returns real-time marker locations with coordinates, data values, and descriptions.

1. **Alerts** (LIVE) — Live forecasts, warnings, and extreme weather events
2. **Standard** (RECENT) — Observation baseline with recent environmental conditions
3. **Thermal** (LIVE) — Global heat mapping and temperature anomalies
4. **Atmospheric** (LIVE) — Air quality, gas dynamics, and pollution monitoring
5. **Hydrological** (RECENT) — Oceanic flux, water events, and flood monitoring
6. **Cryosphere** (ARCHIVE) — Glacial extent, ice sheet conditions, and polar data
7. **Biosphere** (RECENT) — Vegetation health, deforestation, and ecosystem tracking
8. **Starlink** (LIVE) — Real-time satellite constellation position tracking
9. **Orbit Debris** (LIVE) — ESA-sourced space debris tracking and collision risk

## Platform Architecture

QAI Earth is built on a session-authenticated REST architecture.

### AI Query Engine
Natural language processing with Gemini, Google Search grounding for real-time data, automatic location extraction and geocoding.

### Map Visualization
Four map themes (Standard, Satellite, Terrain, 3D Globe), dynamic markers with customizable styles, layer stacking and toggling.

### Data Layer System
Persistent data layers created from queries, custom marker colors/shapes/sizes, bounding box calculation for auto-zoom.

### Report Generation
AI-synthesized reports with executive summaries, key findings, and recommendations. PDF export with embedded map screenshots.

### Collaboration
Share via public links, organization-wide, or with specific users. View and edit permissions with optional expiration.

### Chat Persistence
Multi-turn conversations with message history, version tracking for regeneration, and source citations.

## What kind of questions can I ask?

You can ask about:
- Locations and infrastructure (facilities, landmarks, transportation)
- Earth science (climate, geology, oceanography, meteorology)
- Geographic analysis (population, demographics, urban planning)
- Natural resources (mining, energy, agriculture, water systems)
- Anything that can be visualized on a map

Examples:
- "Show me air quality levels in major European cities"
- "Where are the most earthquake-prone zones in Japan?"
- "What are the largest solar farms in the United States?"
- "Show me deforestation hotspots in the Amazon basin"

## What data sources does QAI Earth use?

QAI Earth uses Google Search grounding for real-time information, combined with data from:
- NASA GISTEMP (surface temperature)
- NOAA Global Monitoring Laboratory (atmospheric gases)
- NASA Sea Level (ocean measurements)
- NOAA NCEI (ocean heat content)
- ESA Space Debris Office (orbital debris tracking)
- World Glacier Monitoring Service (glacial data)
- Global Forest Watch (forest cover)

## FAQ

### What is QAI Earth?
QAI Earth is a spatial intelligence platform that lets you ask natural language questions about any location, climate pattern, or geographic phenomenon. It combines AI-powered analysis with real-time geospatial data from authoritative sources like NASA, NOAA, and ESA, visualized on interactive maps.

### What kind of questions can I ask?
You can ask about locations and infrastructure, earth science topics like climate and geology, geographic analysis including demographics and urban planning, natural resources such as mining and energy, and anything that can be visualized on a map. For example: "Show me air quality levels in major European cities" or "Where are the most earthquake-prone zones in Japan?"

### What data sources does QAI Earth use?
QAI Earth uses Google Search grounding for real-time information, combined with data from NASA GISTEMP, NOAA Global Monitoring Laboratory, ESA Space Debris Office, World Glacier Monitoring Service, and other authoritative scientific sources.

### Can I export my analysis?
Yes. You can export location data as CSV files, generate AI-powered PDF reports with executive summaries and recommendations, and share interactive analyses via public links or within your organization.

### What map views are available?
Four map themes are available: Standard street map, Satellite imagery, Topographic terrain, and a 3D Earth globe with WebGL rendering. You can switch between them at any time.

### Is there an API for QAI Earth?
QAI Earth is built on a REST architecture with endpoints for AI queries, chat management, data layer operations, report generation, and sharing. Access is currently session-based through the web application. Public API key access is on our roadmap — contact us at info@qai.io for enterprise API needs.

### What are system data layers?
System data layers are 9 built-in monitoring feeds that provide continuously updated geospatial data: Alerts, Standard, Thermal, Atmospheric, Hydrological, Cryosphere, Biosphere, Starlink satellite tracking, and Orbit Debris tracking.