Every Feature in Deep Seer

Track 8,000+ active satellites and debris objects in real-time. Each satellite is color-coded by type: military, weather, communications, navigation, science, space stations, and debris. Positions are computed using precise orbital mechanics so what you see on the globe is where the satellite actually is right now.

See thousands of commercial and general aviation aircraft worldwide, updated every few seconds. You can toggle position trails (showing where the plane has been over the last 5, 10, or 15 minutes) and see the planned route from departure to destination drawn as a curved line on the globe.

Track military aircraft globally including tankers, transports, fighters, and surveillance platforms. Each aircraft is identified by type (C-17, F-16, KC-135, etc.) with red silhouette icons facing the direction of travel. Position trails are available via layer settings.

Real-time ship tracking via the Automatic Identification System that all large vessels are required to carry. See cargo ships, tankers, fishing vessels, and naval ships with their heading, speed, and destination port.

Live seismic events from the USGS Earthquake Hazards Program. See where earthquakes are happening right now, with magnitude, depth, and felt reports. Only shows events magnitude 2.5 and above to filter out microseisms.

NEXRAD weather radar composites showing precipitation, storm cells, and severe weather overlaid on the globe. See rain, snow, and thunderstorms in near real-time.

Real-time traffic flow visualization with color-coded congestion levels. At street level, individual vehicles detected from OpenStreetMap data appear with traffic particle animations showing flow direction.

Live traffic camera feeds from 17 sources spanning government DOT agencies and traffic platforms. Supports HLS video streams, auto-refreshing JPEG snapshots, and MJPEG feeds. Covers New York, Los Angeles, Washington DC, Austin, Michigan, Wisconsin, Arizona, Idaho, Utah, Connecticut, Pennsylvania, Georgia, Oregon, Washington State, and more.

Watch up to 9 live camera feeds simultaneously in a configurable grid layout (1x1, 2x2, or 3x3). Each feed shows real-time HLS video or auto-refreshing JPEG snapshots with status indicators showing whether the feed is live.

Ambient air quality monitoring stations worldwide showing PM2.5, PM10, ozone, nitrogen dioxide, sulfur dioxide, and carbon monoxide measurements. Each station is color-coded by Air Quality Index so you can see at a glance where air is healthy (green) or hazardous (purple).

EPA Toxics Release Inventory and ECHO enforcement data showing every factory and facility that releases pollutants into the environment. Includes violation history, penalty amounts, and whether the facility is currently in compliance with environmental laws.

Water quality monitoring stations across US waterways showing contaminant levels, turbidity, dissolved oxygen, and compliance status. See which rivers and streams have clean water and which ones are polluted.

Click any building on the map to identify it by address and type. When you zoom in below 2km altitude, 3D building footprints from OpenStreetMap appear as extruded shapes showing the actual outlines of buildings.

Active fire detections from NASA satellites. Thermal hotspots are plotted on the globe showing where fires are burning right now, updated every few hours with confidence levels indicating how certain the detection is.

Active National Weather Service warnings, watches, and advisories. Tornado warnings, flash flood alerts, hurricane tracks, and winter storm advisories all appear on the globe in real time.

Major disasters worldwide — earthquakes, floods, tropical cyclones, volcanic eruptions — from the Global Disaster Alert and Coordination System. These are the big events that affect entire regions.

USGS streamflow gauges across the US showing water level, discharge rate, and whether a river is at normal, action, or flood stage. Essential for monitoring flood risk in real time.

EPA environmental enforcement actions and OSHA workplace safety violations. See which companies are breaking environmental and worker safety laws, how much they were fined, and whether they are currently in compliance.

Federal contract awards from USASpending.gov. See who is getting government money, how much, from which agency, and where the work is being performed. Essential for tracking public spending.

Federal Election Commission campaign contribution data. See who is donating to which political campaigns, how much they gave, and where the money is going. Covers individual donors and PACs.

Official U.S. Senate Lobbying Disclosure Act filings — over 108,000 lobbying filings showing who is trying to influence government policy, which companies and organizations are paying them, and on what issues. Search by registrant, client, issue area, and year.

Stock trades disclosed by members of Congress under the STOCK Act. See which stocks lawmakers are buying and selling, when they made the trade versus when they disclosed it, the ticker symbols, and estimated dollar amounts.

Shell companies, intermediaries, and officers from the Panama Papers, Paradise Papers, and Pandora Papers leaked by the International Consortium of Investigative Journalists. Search for any person or company to see their offshore connections.

Every NRC-regulated nuclear power plant and research reactor in the United States. Shows operating status, license expiration dates, and exact geographic coordinates.

International weapons flows between countries from the Stockholm International Peace Research Institute. See which countries are selling weapons to whom — supplier, recipient, weapon type, and delivery year.

Crime statistics by state from the FBI Uniform Crime Report and NIBRS data. See violent crime rates, property crime rates, and specific offense types across the United States.

Copernicus Sentinel-2 multispectral satellite imagery for land cover analysis, vegetation health monitoring, and detecting changes over time. See the actual surface of the Earth from space.

NASA imagery showing artificial light emissions from space at night. Useful for tracking urbanization, detecting power outages, monitoring conflict zones, and seeing the actual footprint of human civilization after dark.

Near-real-time deforestation alerts from Global Forest Watch showing where trees are being cut down worldwide. Includes confidence levels and annual tree cover loss statistics.

Two detection modes. "Curated" mode shows known GPS jamming hotspots around the world. "ADS-B Confidence" mode analyzes real-time navigation integrity values from commercial aircraft to crowdsource detection of electronic warfare activity — when multiple aircraft in an area report degraded GPS, it appears on the map.

Temporary Flight Restrictions from the FAA shown as translucent red circles on the globe. Time-aware — only shows restrictions that are currently active based on the timeline position. Auto-refreshes every 15 minutes.

"The hole in the map IS the signal." Detects when expected traffic disappears from an area by comparing current counts against historical baselines. When aircraft or vessels suddenly vanish from a region, it highlights the gap as a red zone — this often precedes military operations, airspace closures, or GPS jamming.

Fragile States Index scores for approximately 180 countries. Measures political, economic, social, and cohesion fragility on a 0-120 scale. Higher scores indicate greater instability and risk of state failure.

World economic data mapped to country capitals — GDP, population, and GDP per capita from the World Bank. See the economic weight of every country at a glance.

Global critical mineral deposits and mines: lithium, cobalt, rare earths, copper, nickel, tungsten, graphite, manganese, platinum, and uranium. See where the minerals that power modern technology come from.

CISA Known Exploited Vulnerabilities catalog mapped to the headquarters of affected software vendors. Shows which companies have active, exploited security vulnerabilities in their products.

Real-time internet connectivity disruptions worldwide. Detects when countries or regions lose internet connectivity — useful for monitoring censorship, infrastructure failures, and conflict-related shutdowns.

Temperature, precipitation, drought, sea ice, sea level, and permafrost anomalies worldwide. Each anomaly is rated by severity so you can see where the climate is deviating most from historical norms.

Geolocated armed conflict events worldwide: battles, violence against civilians, and explosions/remote violence. Each event includes actors involved, fatality counts, and descriptions. Updated in near-real-time.

Geolocated protest and riot events worldwide. See where people are taking to the streets, what they are protesting about, who the actors are, and how events are unfolding.

Global submarine fiber optic cable routes and landing points. These cables carry 99% of the world's intercontinental internet traffic. See the physical backbone of the global internet rendered as polylines on the globe.

Major oil and gas pipeline routes worldwide. See the type (oil/gas), operational status, operator, and which countries each pipeline passes through.

Major hyperscale and AI-capable data centers worldwide — Google, AWS, Azure, Meta, Oracle, xAI, and major colocation facilities. See where the physical infrastructure powering artificial intelligence is located.

Major military installations worldwide including air bases, naval bases, army bases, and joint facilities. Shows branch of service, country, and operational status.

Real-time news monitoring from GDELT covering 100+ countries in 65+ languages, plus RSS feeds from 20+ outlets including Reuters, BBC, NPR, ProPublica, Bellingcat, OCCRP, ACLU, EFF, Amnesty International, and Human Rights Watch.

Monitor Reddit, Bluesky, and Mastodon for OSINT-relevant discussions. Track mentions of specific entities across platforms with geolocation mapping and engagement metrics. Useful for detecting emerging events before they hit mainstream news.

Monitor new regulations, executive orders, and agency rules from the Federal Register. Search by agency or keyword to track regulatory changes that affect your areas of interest.

Aggregate everything known about a person or company across all 90+ data sources into a single intelligence dossier. Cross-references corporate records, sanctions, offshore connections, regulatory violations, government contracts, political contributions, congressional trades, SEC filings, media coverage, and social mentions. Includes name matching that accounts for spelling variations and a composite risk score.

Detects suspicious timing patterns between events. For example: when a member of Congress trades a stock right before an SEC filing about that company, or when a campaign donor receives a government contract shortly after making a contribution. Uses statistical analysis to distinguish genuinely suspicious timing from coincidence.

Maps money and power flows between people, organizations, and government entities as an interactive network graph. See who donated to whom, who lobbied for what, who received contracts, who owns offshore companies. Identifies the most influential nodes in the network and can trace the shortest connection path between any two entities.

Composite risk scores for any entity on a 0-100 scale. Combines four dimensions: regulatory risk (violations, penalties, environmental justice proximity), financial transparency (SEC filings, offshore links, sanctions), political influence (lobbying, contributions, trade volume), and conflict of interest (donor-contractor overlap, lobbyist-regulator overlap).

Automatically detects 9 types of anomalies: aircraft exceeding speed limits, sudden altitude changes, clusters of entities from different data sources converging on one location, signal loss (entities going dark), vessels exceeding speed limits, air quality spikes, military aircraft surges, intense wildfire hotspots, and entities appearing in unusual locations. Uses statistical analysis and machine learning to separate real anomalies from noise.

Six behavioral patterns detected across entities: rendezvous (multiple entities from different sources converging), dark transit (vessels turning off tracking then reappearing somewhere else), loitering near sensitive locations, repeated visits to the same area over time, two entities traveling together, and sudden changes in direction.

Automated cross-domain alert rules that detect meaningful connections between different data layers. Examples: wildfire spreading toward a pipeline, earthquake near a data center, vessel loitering near an undersea cable, GPS jamming affecting aircraft, protests near gunshot detections, deforestation near mineral deposits, offshore entities linked to lobbying activity. Plus 3 multi-step chains: Pollution-to-Health, Regulatory Capture, and Disaster Cascade.

Monitors reporting patterns of tracked entities and flags when something goes unexpectedly silent. If an aircraft, vessel, or ground station suddenly stops transmitting, the system generates an alert showing the pattern of silence across data sources.

Reconstructs the historical path of any tracked entity from discrete position reports. Displays the route as a trail on the globe with automatic detection of stops (where the entity paused) and heading-based path interpolation between reports.

Ask questions in plain English like "show military aircraft near DC in the last hour" and the system translates your query into a structured spatial-temporal search. Supports filtering by entity type, radius, bounding box, and time window.

Enter any location and date/time to find which satellites were overhead at that exact moment. Answers the question: "Was a reconnaissance satellite over this location at 2:00 AM?" Shows satellite name, NORAD ID, altitude, distance, and type (military, communications, weather, etc.).

A panic button that cranks all data polling to maximum speed — flights and military every 10 seconds, earthquakes and weather every 30 seconds — and records everything to the database. Shows your current view and confirms the monitoring area before starting. All captured data is available for timeline replay later.

Selective burst recording — choose which specific data sources to record at maximum speed. Same viewport confirmation as Record Everything, but only captures the feeds you select (for example, flights + military + GPS jamming during a crisis).

All position-based layers buffer 20-30 minutes of historical data. Drag the timeline slider to replay entity movements minute-by-minute on the 3D globe. Satellites use orbital mechanics for any historical time. Burst recording extends the replay window with server-stored positions.

Always-visible voice command panel. Say "Fly to Los Angeles," "show military," "switch to night vision," or ask any question for an AI-powered answer. Uses your browser's built-in speech recognition (Chrome and Edge). A collapsible hint card shows available commands.

Automatically extracts people, organizations, places, dates, and monetary amounts from unstructured text. Enriches news articles and social media posts for better cross-referencing across data sources. Works with a lightweight local library or an optional local AI model for higher accuracy.

Upload PDFs, text documents, court filings, and memos for automated analysis. The system extracts entities, maps relationships between them, and detects events mentioned in the text. Combines pattern matching with AI analysis for comprehensive results.

Claude-powered analysis capabilities: contradiction detection (comparing official statements vs. documented facts), hypothesis generation, narrative pattern analysis, credibility assessment, evidence triage (classifying urgency and relevance), and translation with entity extraction. Uses cost-efficient AI models for simple tasks and more capable models for complex reasoning.

Generate comprehensive reports in HTML, Markdown, and printable formats. Six templates: Entity Profile, Incident Timeline, Accountability Audit, Influence Map, Investigation Summary, and Document Analysis. Each report is publication-ready.

Every piece of evidence is cryptographically signed to prevent tampering. Each item gets a unique fingerprint (hash), a server signature proving it was not altered, and an optional trusted timestamp proving when it was captured. Verification checks the math to confirm the evidence has not been changed.

Verify where a photo or video was actually taken using three methods: landmark matching (comparing reference points in the image to known landmarks), shadow analysis (calculating when a shadow could have been cast based on sun position), and building footprint comparison. Results can be saved directly to the evidence archive.

Finds geographic clusters of incidents, detects whether incident rates are escalating over time, tests whether attacks disproportionately occur near protected sites (hospitals, schools, places of worship), and links military unit positions to nearby incident locations.

Set up alerts for specific entities — people, vessels, aircraft, or patterns — and get notified the moment they appear in any data source. Supports exact names, vessel IMO numbers, aircraft tail numbers, radio hex codes, and even regex patterns. Fuzzy matching catches spelling variations.

Real-time plausibility checks on all incoming data. Detects AIS spoofing (ships moving impossibly fast), fake flight data (aircraft climbing faster than physics allows), duplicate entities, and disagreements between data sources. Monitors the health of every data source with error rates, latency, and degradation detection.

Analyze how resilient infrastructure networks are to disruption. Pick a network (submarine cables, pipelines, flight routes, supply chains) and see which nodes are most critical. Simulate what happens when key nodes fail — watch the cascade of failures propagate through the network in stages.

See where your data coverage is strong and where the blind spots are. The globe is divided into hexagonal zones, each scored by how many different data sources are reporting from that area. Red zones have poor coverage; green zones have good multi-source verification.

Automatically checks news articles and social media posts against all 90+ data sources. When a tracked entity appears in media coverage, the system flags it. Detects surges in media attention and correlates coverage spikes with other events.

Track media tone and sentiment over time for any entity. See whether coverage is becoming more positive or negative, detect sudden shifts in public perception, and identify coverage spikes with statistical baselines.

Predicts where contamination will travel downstream from a spill point. Uses the actual stream network and real-time river flow data to estimate when pollution will reach each downstream community.

Log incidents (airstrikes, shelling, civilian harm, and 6 other types) with precise geographic markers via Crisis Mode. Each incident is stored permanently with severity rating, stage tracking, and pattern flags. View all incidents in the Investigation Panel.

A six-stage checklist based on the Berkeley Protocol for digital open source investigations: Discovery, Collection, Preservation, Verification, Analysis, and Reporting. Each stage has 3 items. Completion score tracks your progress (0-100%). Stages are advisory — they never block your work.

Generate court-ready evidence packages formatted for specific jurisdictions: International Criminal Court (Rome Statute), European Court of Human Rights (Rule 47), US Federal Courts (FRE 901/902), or generic export. Includes chain-of-custody appendix, hash verification table, and audit trail.

Unified screening panel with three tabs: Adverse Media (searches news archives for negative coverage with risk scoring), Sanctions Screening (checks 100+ global watchlists for matches), and Identity Resolution (cross-references FEC, SEC, OpenCorporates, Senate LDA, ICIJ, and ProPublica records).

Analyzes historical data — earthquakes, armed conflicts, protests, anomaly clusters — to predict where incidents are likely to occur next. Generates scored predictions with contributing factors, timeframe estimates, and confidence levels. Off by default.

Predicts what a moving entity (aircraft, vessel, vehicle) will do next based on its position history and behavioral patterns. Classifies behavior as routine, transit, loitering, evasive, or surveillance. Shows predicted actions with probability bars and reasoning. Requires manual activation per entity.

Every analysis result records its provenance — which data sources it came from, what methods were used, and how confident the result is. You can trace any conclusion back through the complete chain of reasoning to the raw data.

Archive any web page as evidence. The system fetches the full HTML, extracts metadata (title, author, publication date), and stores the complete page with tamper-proof hashing. Duplicate pages are automatically detected.

Capture and preserve social media posts from Twitter, YouTube, Reddit, Bluesky, Telegram, and Facebook. Stores the embedded content, author information, post ID, and platform metadata as evidence.

Upload up to 20 files at once. Photos have their camera information and GPS coordinates automatically extracted. All files receive the full evidence integrity treatment — hashing, signing, and timestamping.

Receive evidence packages from anonymous whistleblowing platforms (SecureDrop or GlobaLeaks) via a secure webhook. Automatically creates evidence records with cryptographic integrity checks.

Track sources by pseudonyms — never real names. Contact notes are encrypted so that even if the database is accessed, source identities remain protected. Includes credibility scoring and submission tracking.

Removes camera information and location data from photos in your browser before they leave your computer. Protects source identity even if the server is compromised. Toggle between stripped mode (source protection) and preserved mode (verification).

Translate text between 30+ languages with entity extraction. The system identifies people, organizations, places, dates, and monetary amounts in the translated text. Translation history is stored for reference. Essential for foreign-language evidence.

Define your own entity types (with custom names, colors, icons, and properties) and relationship types. Create the data model that fits your investigation. Import and export definitions as JSON for team sharing.

Per-representative scorecards showing how long they wait to disclose trades, which stocks they concentrate in, their buy/sell ratio, and an overall suspicion score (0-100). Flags trades that precede regulatory events and committee-relevant sector concentration.

Cross-references campaign contributions with government contract awards. Flags cases where a donor (or their employer) receives a federal contract within 180 days of making a political contribution. The pay-to-play detector.

Matches entities from the Panama Papers, Paradise Papers, and Pandora Papers against domestic corporate records and EPA facility operators. Surfaces potential shell company ownership of polluting facilities or government contract recipients.

Identifies companies that are actively violating environmental or worker safety laws while simultaneously receiving taxpayer-funded government contracts. An accountability flag for procurement oversight.

Correlates lobbying expenditures with subsequent government contract awards to the same or related entities. Maps the influence pipeline from lobbying registration to contract receipt. Temporal analysis detects suspicious timing.

AI-generated intelligence briefings for any location on the globe. Click anywhere and get a coherent narrative summary synthesizing nearby entities, environmental conditions, recent events, accountability flags, and relevant data from all active layers.

County-level mortality data for cancer, respiratory disease, cardiovascular disease, and kidney disease from the CDC, combined with Social Vulnerability Index data. Color-coded by age-adjusted death rate relative to the national average. See which communities bear the heaviest health burden.

Cross-references what facilities are permitted to discharge versus what they actually release. Scores facilities by how far they exceed their limits, how long they have been out of compliance, and their Significant Non-Compliance flags. Surfaces the worst offenders.

Traces pollution from source to impact as a visual chain: facility permits lead to actual releases, which are carried by wind and water to downstream communities (weighted by environmental justice vulnerability), which suffer health outcomes tracked by the CDC. Each link shows a confidence score.

Move beyond correlation to establish causation. Four rigorous statistical methods: Granger causality (does A predict B over time?), Difference-in-Differences (what was the impact of an event?), Regression Discontinuity (what happens at a threshold?), and Synthetic Control (what would have happened without the intervention?). All computed from scratch with no external libraries.

Monitors proposed and final rules from the federal regulatory system. Classifies each rule as weakened, strengthened, or procedural. Cross-references with lobbying data to detect regulatory capture patterns — when industry lobbying leads to weaker regulations.

Tracks Freedom of Information Act requests via MuckRock. Detects overdue requests (30-day threshold), normalizes statuses across agencies, and scores each agency's compliance (0-100) based on how often they miss deadlines and how long they take to respond.

Traces corporate ownership chains up to 4 levels deep. Uses SEC filings, OpenCorporates, and ICIJ Offshore Leaks to find who really owns a company. Detects shell companies by scoring jurisdiction risk, generic name patterns, and nominee directors. Overall opacity score measures how hidden the ownership structure is.

Click "Ownership" on any facility to see polyline arcs on the globe connecting all facilities owned by the same parent company. Subsidiaries, officers, and beneficial owners shown as color-coded arcs. Shell companies and offshore entities highlighted in red with dashed lines.

Compares actual environmental readings against EPA/NAAQS standards to show the "counterfactual gap" — how much worse the air or water is compared to what the law says it should be. Identifies which nearby facilities are most responsible using distance-weighted scoring.

On-demand historical trend analysis. Fits a line to historical data to show whether things are getting better or worse, detects inflection points where trends changed, and correlates different metrics against each other. AI-generated narrative summaries explain the findings.

Stamps evidence hashes to the Bitcoin blockchain via public OpenTimestamps calendars. Provides mathematically provable proof-of-existence that cannot be backdated — even by the operator of Deep Seer. Proofs automatically upgrade when the Bitcoin block confirms.

Six pre-built investigation workflows: Environmental Contamination, Campaign Finance, Corporate Accountability, Government Contract Fraud, Systemic Inequality, and Police Accountability. Each template has 5 guided steps, recommends specific layers and panels, and includes a data collection checklist.

Search federal court records via the CourtListener RECAP archive. Look up parties by name, view docket details, and track case status. Identifies litigation patterns useful for corporate accountability investigations.

FEMA National Shelter System showing evacuation shelters across the US with capacity, wheelchair accessibility, and operational status. Know where the nearest shelter is before you need it.

Calculated blast radius overlays for approximately 30 major US nuclear target sites. Three concentric rings show the fireball (1 mile), severe blast damage (5 miles), and moderate damage zone (10 miles).

EPA SEMS contaminated sites requiring long-term cleanup. Color-coded by NPL (National Priorities List) status: active sites in red, proposed in orange, and cleaned sites in green. These are among the most polluted places in America.

US counties where people must drive more than 30 minutes to reach the nearest emergency room. Highlights the accelerating wave of rural hospital closures, especially in states that did not expand Medicaid.

State-level maternal death rates per 100,000 live births with racial disparity ratios. The US has the worst maternal mortality rate in the developed world — and it varies 4x between states. Black women die at 2-3x the rate of white women.

State-level drug overdose mortality rates with fentanyl percentage breakdown. Maps the overdose crisis across America — see which states are losing the most people and how much of the epidemic is driven by fentanyl.

USDA Food Access Research Atlas showing census tracts where residents cannot easily buy fresh, affordable groceries. Food deserts strongly correlate with poverty, race, and poor health outcomes.

State-level lead poisoning risk from pre-1978 housing (when lead paint was banned), lead water service lines, and child blood lead levels. Lead causes permanent brain damage in children — there is no safe level of exposure.

Per-pupil education spending by state (ranging from $8,500 to $28,000), graduation rates, Title I percentage, teacher shortages, and the funding gap between rich and poor districts within each state.

Census Gini coefficient by state — an income inequality index from 0 (perfect equality) to 1 (one person has everything). Includes median income and poverty rate context. Washington DC and New York lead the nation in inequality.

The gap between what workers are legally required to be paid (state minimum wage) and what they actually need to survive (MIT Living Wage calculation for a single adult). Many states have $7-10/hour shortfalls.

Department of Labor Wage and Hour Division enforcement data: how much stolen wages were recovered and how many employees were affected. The actual amount of wage theft is estimated to be 3-10x what enforcement catches.

Workplace safety enforcement data: serious and fatal violations, penalty amounts, inspection counts, and workplace deaths by state. See where workers are most at risk on the job.

Princeton Eviction Lab data showing eviction filings, completed evictions, and renter demographics. Southern states lead the nation in eviction rates. Eviction is both a cause and consequence of poverty.

Which states pay more in federal taxes than they receive back (donor states, shown in blue) versus which states receive more than they pay (recipient states, shown in red). Challenges the "maker vs taker" political narrative with actual IRS and spending data.

Institute for Justice grades (A through F) for civil asset forfeiture protections by state. Shows where police can seize cash, cars, and homes without ever convicting — or even charging — the owner with a crime.

Polling place closures since the Supreme Court gutted the Voting Rights Act in Shelby County v. Holder (2013). Texas alone closed 750+ polling locations. Shows per-capita availability and average wait times.

Efficiency gap, compactness scores, number of competitive districts, and redistricting method for each state. Shows which political party benefits from how the electoral map was drawn.

Local journalism is dying — over 2,900 newspapers have closed since 2005. This layer shows counties with zero local news coverage, where government corruption and corporate malfeasance can go unchecked because nobody is watching.

Mass incarceration mapped. Prison and jail populations, racial disparity ratios (often 5-6x), cost per inmate ($30,000-60,000 per year), and percentage of prisoners held in private, for-profit facilities.

Every ICE detention facility in the country — many operated by private prison companies CoreCivic and GEO Group. Shows capacity, daily population, average detention length, and reported deaths.

Police killings per million population, percentage of victims who were unarmed, racial disparity in police violence, and officer accountability rates (rarely above 2%). This is data the federal government refuses to systematically track.

Where America lacks internet access. State-level broadband gap data from the FCC — Mississippi has roughly 25% of residents without broadband, while Northeastern states average about 4%. The digital divide mapped.

Public transit accessibility scores for 50 US metros on a 0-100 scale. NYC scores 90, while many Sun Belt cities score below 30 — forcing low-income residents into car dependency with few transportation options.

A full 3D model of the Earth with Google Photorealistic 3D Tiles providing photogrammetric terrain and building models. Full camera control with fly-to animations, terrain clamping, atmospheric rendering, and day/night lighting.

A custom visual filter that makes the globe look like an old CRT monitor — complete with phosphor scanlines, barrel distortion, color bleeding, and screen curvature. Retro surveillance aesthetic.

True night vision goggles simulation. Terrain is crushed to near-black while entity icons and markers glow bright phosphor green. Steady glow with subtle static grain and tight vignette — no annoying flashing.

False-color thermal imaging simulation. Maps scene brightness to an ironbow palette (black-purple-red-orange-yellow-white), creating the appearance of a forward-looking infrared camera.

Falling binary code rain with terrain classification — ocean, land, and buildings render as bright neon green code columns. Building footprints from OpenStreetMap accumulate persistent green tiles.

Studio Ghibli-inspired cel-shading with ink outlines. The globe and entities are rendered with limited color palettes and soft gradient banding, creating an illustrated look.

CCTV-style rendering with desaturated colors, a timestamp overlay, recording indicator, and subtle scan lines. Makes the globe look like a security camera feed.

Warm false-color heat visualization for analytical overlays. Terrain is desaturated and data density is remapped through an orange-red gradient. Bright hotspots bloom with warm glow. Designed for pollution, health outcome, and violation density analysis.

Cool blue-silver chronographic mode for temporal analysis. A horizontal scan line sweeps the scene. Entity trails show movement ghosting. Pairs with time slider controls for historical scrubbing.

Red-amber threat detection overlay. Areas where lobbying, regulatory weakening, health outcomes, and violations overlap glow progressively brighter. The more risk factors converge on one location, the more intense the visual signal.

Corporate identity visualization. Blue-gray terrain with boosted ownership markers. Facilities are color-coded by their ultimate parent company, revealing which corporation controls what.

Compliance traffic-light system. Green areas are in compliance, yellow are marginal, red are in violation. Sterile government-document aesthetic with red pulses on violation hotspots.

Environmental justice vulnerability gradient from purple to red. Combines EJScreen environmental burden scores with Social Vulnerability Index demographic data. Warm, humanizing palette with pressure-wave edge glow on the most vulnerable communities.

Tactical amber shader with desaturation, warm tint, breathing pulse border, crosshair markers, and film grain. Includes a full crisis workspace: HUD status bar, floating action toolbar, incident logging directly on the globe, viewport evidence capture, and Berkeley Protocol verification workflow. Press C to activate.

Swap the 3D photorealistic globe for a flat OpenStreetMap tile layer. The View selector is independent of the Mode (shader) selector, so you can use any analytical mode with either the 3D globe or the flat map.

Purpose-designed icons for every entity type. Red military aircraft silhouettes with type labels, blue commercial flight silhouettes with callsigns, CCTV camera thumbnails with FOV cones, individual vehicle dots with speed indicators, and more. All icons orient to heading and scale with zoom.

3D building outlines from OpenStreetMap appear when you zoom below 2km. Click any building to identify it by address and type via reverse geocoding.

Corner bracket frames drawn around every point entity across 13 data layers. Gives everything the visual aesthetic of a military tracking/targeting system.

Entities on the far side of the Earth are hidden at all zoom levels. Prevents visual clutter from things you cannot actually see from your current vantage point.

Downwind pollution plumes from factories (accounting for actual wind direction), seismic shaking intensity rings from earthquakes, and radial influence zones around any point entity. See the area of effect, not just the point.

Create named investigations. Pin entities, add annotations, attach evidence (notes, links, news articles, social posts, documents), and auto-generate leads. Each investigation has its own graph view, timeline, and export options (Markdown and JSON).

Force-directed node graph showing relationships between pinned entities in your investigation. Automatically discovers related entities from the intelligence pipeline. Click any node to drill into its full dossier.

Search across all data sources with type filter chips, color-coded results, keyboard navigation (arrow keys + Enter), result count, and recent search history. Separate "fly to" and "select" actions for each result.

Select 2-4 entities and view their properties side-by-side in a comparison table. Differences between entities are highlighted in amber for easy identification.

Faceted search and filter panel for browsing ALL entities across all active layers. Filter by entity type, data source, time range, and region. Includes pie charts, bar charts, a 24-hour activity histogram, and a virtual-scrolled result list. Click any entity to fly to it.

Save any structured or natural language query to re-run later. Queries persist to the database when an investigation is active (localStorage fallback otherwise). Paginated results with 20 per page.

Draw circle geofences on the globe with lat/lon center and radius. Filter by entity type (vessel, aircraft, military, satellite). Toggle active/inactive. Get real-time alerts when entities enter or exit the fenced area.

Save the current camera position, active layers, visual filter, and time range as a named view. Restore any saved view with one click to return to exactly where you were.

Measure distance between two points and area of a polygon directly on the globe. Uses geodesic (great circle) calculations for accurate measurements on a spherical surface.

Build guided tours of the globe with annotated waypoints. Auto-play flies the camera through each location with 5-second dwell time. Export the completed story as a standalone HTML file that anyone can open.

Build custom dashboards with live metric cards. Choose from entity count, vessel traffic, air quality trend, and violation count metrics. Each card shows a sparkline trend chart that auto-updates every 30 seconds. Drag cards to reorder.

Export the current view as a screenshot, entity data as CSV, investigation notes as Markdown, or raw data as JSON. Everything you see can be exported for reporting and publication.

Quick keys: / opens search, ? shows help, L toggles layers, Escape deselects, Space plays/pauses timeline, F toggles fullscreen. All shortcuts are disabled when you are typing in a text field.

Every AI-generated result (chat responses, briefings, threat predictions, correlations, anomalies, translations, screenings, agent outputs) is automatically saved to a persistent log. Star, tag, annotate, link to investigations, compare side-by-side, and export as JSON.

Four graph visualization tools: spatial entity networks by location, interactive search with depth 1-4 expansion, pinned investigation entity relationships, and lobbying/influence network mapping. All color-coded by node type (person, company, facility, government) with labeled edges.

Dual synchronized globe views for temporal comparison. Each pane has its own date/time picker — set one to "now" and the other to "6 months ago" to compare the same area at different points in time. Cameras stay synced so both views show the same location.

Compare entity positions and density between two different time periods. Side-by-side temporal snapshots reveal changes in patterns over days, weeks, or months.

Snapshot all visible entities at any moment, then take another snapshot later. The system shows which entities appeared, disappeared, or changed properties between the two captures — a diff for the physical world.

Reveals how relationships between variables change across geography. For example, the relationship between air pollution and health outcomes might be strong in one region and weak in another. Includes spatial autocorrelation testing (Moran's I) and local clustering analysis (LISA). Preset analyses for pollution-health, income-air quality, and regulatory compliance.

Predicts future values using two established methods: Holt-Winters (which captures trends and seasonal patterns) and ARIMA (which captures autocorrelation). Combines both into an ensemble forecast with widened prediction intervals. Includes threshold breach estimation (when will a value exceed a limit?) and backtesting to measure accuracy.

Structured hypothesis testing with falsification criteria and Analysis of Competing Hypotheses (ACH) — a method used by intelligence analysts to evaluate multiple explanations against evidence. The system suggests automated tests and tracks which hypotheses are supported or contradicted.

Monitors your analysis behavior for cognitive biases: confirmation bias (only looking at supporting evidence), anchoring (over-relying on first data point), availability bias (overweighting recent or vivid events), base rate neglect, and premature closure. Produces an investigation health score.

Record, parameterize, and replay analysis workflows so anyone can reproduce your methodology. Fork and modify existing workflows. Export methodology for peer review. Includes 15 built-in presets: Area Recon, Entity Deep Dive, Change Over Time, Threat Proximity, Source Cross-Validation, Illegal Dumping, Dark Vessel, Follow the Money, Protest Escalation, Environmental Justice, Supply Chain Disruption, CCTV Anomaly, Civic Health Audit, Criminal Justice System Audit, and Nuclear Preparedness.

JARVIS-style conversational AI. Ask questions in natural language and the AI can control globe layers, fly to locations, search entities, run analyses, and generate briefings. Supports voice input and optional text-to-speech output. Chat history persisted locally. Save any response to an investigation.

Real-time interrupt alerts when significant events occur. Listens to all alert channels — anomalies, correlations, environmental alerts, watchlist matches, and signal gaps. Audio and visual notifications with severity-coded cards. Save alerts for review and export as JSON. Off by default.

Two tools in one panel. AI Agent Builder lets you chain LLM calls, queries, correlation checks, layer toggles, fly-to commands, and alerts with 7 trigger modes. Pipeline Builder is a visual step-based editor for Data Source, Filter, Transform, Analyze, and Output steps.

Register AI agents to run on automated schedules (every 5 minutes, 15 minutes, 1 hour, or daily). The server executes agent steps on the configured interval. Toggle agents on and off, view status and results via the management API.

Monitor multiple camera feeds simultaneously in a 2x2 or 3x3 grid. Auto-scan mode rotates through available cameras every 10 seconds. Motion alert mode fires notifications when motion exceeds a threshold. Save snapshots from any feed.

Click anywhere on the globe to see precise lat/lon (6 decimal places) in a bottom overlay. Copy as decimal or DMS format. Manually enter coordinates in any format and fly there. Quick-action buttons to get a contextual briefing or query nearby entities.

Mouse over any entity on the globe to see its name and type in a floating tooltip. Works for flights, vessels, satellites, facilities, earthquakes, and more.

A collapsible legend overlay in the bottom-left shows what colors mean for all active layers. Supports both discrete dot legends and gradient bars.

Centralized visual design tokens for colors, spacing, typography, shadows, transitions, and panel sizes shared across 49 panels. CSS Modules on the top 10 most-used panels. Full ARIA accessibility.

Right-click any entity or location on the globe for a context menu with 7 actions: select entity, pin to investigation, fly to location, copy coordinates, search nearby, ask AI about it, and save the current view.

Press Ctrl+Z to undo your last action, Ctrl+Shift+Z to redo. Tracks up to 50 actions including layer toggles. Automatically skips when you're typing in a text field.

Toggle between dark mode (default, designed for low-light analysis sessions) and light mode (high-contrast for daytime use). Your preference is remembered across sessions.

Dock any panel to the left, right, or bottom of the screen — like VS Code or a professional trading terminal. Docked panels show as tabs with resize handles on the edges. Undock with one click to return to floating.

Spreadsheet-style table view of all entities. Sort by any column, filter across all data, toggle column visibility, select rows with checkboxes, and export filtered results as a CSV file. Handles up to 5,000 entities with pagination.

Save your current panel arrangement as a named layout — which panels are open, where they're positioned, which layers are enabled, and which visual filter is active. Switch between saved layouts instantly.

Panels and UI elements use spring-based animations for a natural, snappy feel — panels overshoot slightly then settle into place. List items cascade in with staggered timing. All animations respect your system's reduced motion setting.

A mandatory responsible use notice shown the first time you visit. Explains that all data is publicly available, prohibits stalking and harassment, notes that face detection runs only in your browser, and requires you to acknowledge before entering the app.

All AI-generated content — chat messages, briefings, predictions, analyses — automatically syncs to the server. If you clear your browser or switch devices, your AI history loads from the server on startup.

Daily AI token limits protect against runaway costs: Free tier gets 50K tokens/day (hard cap), Pro gets 500K (soft cap with $10/1M overage), Enterprise gets 5M (soft cap with $5/1M overage). Check your usage anytime via the API.

Reads license plates from CCTV camera frames using optical character recognition. Validates plate formats, extracts bounding boxes, and automatically checks plates against your watchlist — alerting if a match is found.

Real-time data push via WebSocket for instant entity updates. Free tier uses REST polling (slightly delayed); Pro gets sub-second push for all entity types.

All entities are indexed in a spatial data structure that makes "find everything within 10km of this point" queries nearly instantaneous, even with 100,000+ entities loaded.

Install Deep Seer as a native-feeling app on your desktop or phone. Works offline for cached data. Auto-updates silently when new versions are deployed.

Capture evidence even without internet — screenshots, notes, and incident markers are stored locally and automatically synced when connectivity returns. Investigation data is cached for offline reading.

Invite team members to investigations with role-based access (owner, analyst, reviewer, viewer). See who else is working in real time with presence indicators and a live activity feed.

Control where your evidence is physically stored. Per-investigation region overrides, transfer restrictions for EU compliance, and a visual dashboard showing where your data lives.

Defense-in-depth security: Content Security Policy headers, CORS origin allowlist, CSRF protection, WebSocket rate limiting (5 connections per IP), input validation on every endpoint, bcrypt password hashing, JWT refresh token rotation with revocation detection, and heartbeat-based dead connection cleanup.

Automatic retry with exponential backoff and jitter. Per-adapter circuit breakers that stop calling a failing API after 5 errors and retry after 5 minutes. Client-side WebSocket reconnection with buffered messages. App-level error boundary with Sentry crash reporting.

Structured logging, performance metrics, error tracking, and pre-built monitoring dashboards. Automated data cleanup and database backup scripts included.

Free (30 requests/minute, all data layers), Pro ($9/month, 120 req/min, intelligence + export), Enterprise ($29/month, 600 req/min, split-screen + network graph). Stripe integration for payment processing.

Server-side adapters for 70+ distinct data sources. Each adapter handles API authentication, pagination, rate limiting, data normalization, and entity extraction. Includes 24 equity/accountability adapters covering FEMA, EPA, CDC, USDA, Census, DOL, OSHA, FCC, CMS, and more.

All users, investigations, evidence, documents, and reports persist across server restarts. 14-model database schema with WORM (Write Once Read Many) evidence protection, full-text search, content-addressable storage, and RFC 3161 timestamping.

Comprehensive automated testing: 170+ unit and integration tests across 25+ test files covering authentication, middleware, services, adapters, and client stores. End-to-end browser tests with Playwright. Coverage reporting in CI.

Full responsive design. Panels become bottom sheets on mobile. Touch-friendly 36px tap targets. Landscape optimization. Print styles. Breakpoints for mobile, tablet, and large desktop.

Full accessibility compliance: 4.5:1+ color contrast, focus trapping in every modal panel, Escape-to-close on all overlays, screen reader announcements, reduced-motion support, high contrast mode (Alt+H), and tooltip hints on all icon buttons.

Multi-language support with browser language detection. English, Spanish, and French translations built in. Extensible locale system for adding more languages.

When zoomed out, thousands of entities are aggregated into numbered clusters for performance and readability. Zoom in to expand clusters into individual entities. Automatic re-clustering when the camera moves.

Automatically detects your device capabilities (CPU, GPU, RAM, screen, network) and adjusts performance accordingly. Low-end devices get fewer simultaneous layers and lower resolution. High-end devices get the full experience.

Five zoom presets (Global, Country, State, Metro, City) automatically show/hide layers appropriate for the current altitude. CCTV and buildings only appear at city zoom; satellites and nightlights are visible globally.

Multi-step welcome overlay for new users: welcome screen, recommended starter layer picker (6 layers with toggle pills), and quick-start tips. Pre-selects satellites, flights, and earthquakes. Empty-state hints when no layers are active.

The same web app wrapped in a native shell for iOS (App Store) and Android (Play Store) distribution. Includes splash screen, status bar integration, and push notification support.

When the same entity appears under different names across data sources, the system matches them using fuzzy name matching and phonetic comparison. All aliases are merged together so you see one entity instead of duplicates.

Dual-key management with automatic rotation. Google Maps keys auto-rotate via Cloud API. Background scheduler warns at 60/90/120 days. Admin endpoints for manual rotation and emergency revocation.

Photorealistic 3D tiles, geocoding, and places API. Google provides $200/month in free credit. Light usage stays within the free tier. Beyond that, pay-per-use pricing applies.

25,000+ traffic cameras across 200+ US cities with redistribution rights. Contact sales for pricing (estimated $500-2,000/month).

On-demand satellite tasking for high-resolution imagery of any location on Earth. $90-500 per scene depending on resolution and urgency.

Synthetic Aperture Radar imagery — sees through clouds and at night. Useful for construction detection, military movements, and flood mapping. Approximately $1,000 per scene.

Screens entities against global sanctions lists, politically exposed persons (PEPs), and criminal watchlists. Free for non-commercial use (30-day trial with business email). Commercial use requires a paid license.

Corporate ownership networks, directorships, and company filings across 200M+ companies in 140+ jurisdictions. Free for open data projects, academics, NGOs, and journalists. Commercial use requires a paid license.

Pop any panel into its own browser window for multi-monitor setups. The detached panel stays fully live and synced with the main application. Closing the popup reattaches the panel. Available on AI Chat, Investigations, Dashboard, and AI Outputs.

Historical flight and drone tracks rendered as 3D lines with actual altitude in the globe. Color-coded by altitude change: green for climbing, yellow for level flight, red for descending. Visualize trajectory history, altitude changes, and loiter patterns in 3D space.

Auto-generated camera tours of investigation areas. The camera chains fly-to animations to create narrated fly-throughs of key locations, entities, and events. Playback controls let you adjust speed and pause.

Detect people, vehicles, and 80 object categories in camera feeds using machine learning that runs entirely in your browser — no data is sent to external servers. Bounding box overlays with configurable confidence thresholds.

Cross-feed face matching using face detection and embedding comparison. All processing runs in your browser — no face data is ever sent to servers. Requires explicit user consent before activation. Full ethics gate built in.

Live drone telemetry from DJI FlightHub 2 and Skydio Cloud. Fleet status dashboard (flying/idle/offline), individual drone tracking with battery and signal, and mission waypoints on the globe. Requires enterprise API access plus physical drone hardware.

Software-defined radio for spectrum monitoring. Supports RTL-SDR (~$25), HackRF One (~$350), and Airspy R2 (~$200). Waterfall spectrum display, signal detection with modulation classification, and geolocation via triangulation. Requires a physical SDR dongle connected to your computer.

Full gunshot detection pipeline supporting SoundThinking (ShotSpotter), Shotpoint Acoustic, and custom sensor arrays. TDOA triangulation locates shots within approximately 25 meters. Waveform classification distinguishes gunshots from fireworks and car backfires. Requires deployed sensor hardware or enterprise API.

On-demand satellite redirect to capture imagery of a specific location. Planet Labs SkySat ($90-500/scene), Maxar SecureWatch (enterprise pricing), Airbus Pleiades (sub-meter, ~$10/km2). Requires paid commercial satellite API subscription.

Synthetic Aperture Radar sees through clouds, at night, and detects ground changes. Capella Space and ICEYE offer commercial tasking. Useful for construction, military movements, and flood mapping. Approximately $1,000 per scene.

Holographic data visualization using WebXR. View the intelligence globe in augmented reality through a headset or phone camera. Requires an AR-capable headset (Quest, HoloLens) or modern smartphone.

IRS Form 990 data for nonprofits — revenue, expenses, executive compensation, and mission statements. ProPublica shut down their free Nonprofit Explorer API.

Lobbying registrations, expenditures, and PAC contributions. OpenSecrets restricted public API access — data now requires institutional partnerships or paid access. Replaced by Senate LDA direct integration.

State-level campaign finance data from the National Institute on Money in Politics. Tracked contributions, expenditures, and donors across all 50 states. Service has been discontinued.