If you’ve seen a screenshot of a glowing world map tracking planes, earthquakes, and news all at once and wondered what it was, that’s probably Osiris AI. Osiris is a free, open-source real-time global map dashboard that combines live flight tracking, CCTV feeds, earthquake data, and global news into a single view, run locally on your own computer rather than through a hosted app or website you sign up for.
It’s been popping up in screenshots and short clips online lately, often described as a “Palantir alternative,” and that comparison is doing a lot of the confusion. This page covers what it actually is, what data it pulls and from where, whether Osiris AI live demos are genuinely real-time, whether it’s free, how to actually run it yourself, how its built-in AI feature works, and what it’s realistically meant for, since the honest answer sits well short of the movie-style surveillance tool the Palantir comparison implies.
What Is Osiris AI Used For?
Osiris is what’s called an open-source intelligence platform, OSINT for short, meaning it pulls together information that’s already public rather than accessing anything private or restricted. In practice, that means it’s aimed at journalists tracking a developing story, researchers studying global events, and hobbyists who just like watching the world happen in real time, not exclusively security professionals, even though the underlying technology overlaps with tools used in that field.
The actual experience is closer to a live global newsroom wall than a spy thriller. You open it in a browser, and a dark, GPU-rendered WebGL map fills with moving flight icons, pulsing earthquake markers, and clickable news dots, each one a live layer you can toggle on or off depending on what you’re actually interested in following at that moment. Under the hood, most active forks are built on Next.js and React with MapLibre GL for the map rendering, which is what lets the dashboard smoothly display thousands of moving points at once without the browser choking, something a simpler map implementation would struggle with at this scale. As a situational awareness tool, its whole value is in that single-glance overview, seeing multiple kinds of live global activity at once rather than checking five separate websites.
What Data Does the Osiris World Monitor Show?
Different forks of the project vary somewhat in exactly which layers they include, but a core set of data shows up consistently across the active versions.
Live Flight Tracking
Aircraft positions come from the OpenSky Network, a real, established open flight-tracking data source used well beyond this project, covering commercial, private, and in some forks, military aircraft movements. This works without any account or API key for basic use, though a couple of forks let you optionally add OpenSky credentials for slightly more complete data.
CCTV Camera Networks
This is the detail most worth clarifying directly, since it’s the one people misread fastest: the camera feeds Osiris pulls into its CCTV network map come from public, government-operated traffic camera systems, sources like TfL (Transport for London), WSDOT (Washington State), Caltrans (California), NYC DOT, and VicRoads (Victoria, Australia) are named directly in the project’s own documentation. These are the same public traffic cameras many of these agencies already publish on their own transportation websites, not private security footage and not anything Osiris is accessing without authorization. Some of the more feature-complete forks report aggregating over 2,000 of these public cameras into one browsable map.
Earthquake and Disaster Monitoring
As an earthquake monitoring tool, seismic data comes from the USGS (United States Geological Survey) earthquake API, typically filtered to magnitude 2.5 and above, updating in near real time as new events are detected. Several forks extend this further with wildfire detection from NASA’s FIRMS fire-mapping system and broader disaster tracking through GDACS and EONET, giving a wider view of active natural disasters beyond just earthquakes.
Conflict Zone Mapping
Conflict zone layers plot reported activity in active geopolitical hotspots, generally sourced from aggregated open news and event-tracking feeds rather than any classified or restricted source. Rather than pulling from a single dedicated conflict-tracking API the way flight or earthquake data does, this layer tends to be built by synthesizing signals across the news feeds and event data the dashboard already ingests, which is also why accuracy here leans more on how recently a fork’s news pipeline was updated than the other, more directly sourced layers.
This is also the layer the built-in AI analyst feature, covered further down, most often gets used alongside, cross-referencing flight activity against these zones to answer a question like whether unusual aircraft movement is happening near a currently active hotspot.
Maritime and Vessel Tracking
Several of the more feature-complete forks also include a maritime layer, tracking vessel positions across major global shipping routes using AIS (Automatic Identification System) data, the same kind of tracking that dedicated ship-tracking sites rely on. Coverage in these forks reportedly spans dozens of major global ports and roughly ten key maritime chokepoints, the narrow shipping lanes, like straits and canals, where global trade traffic concentrates. Some forks render this with directional arrows showing each vessel’s heading and type, styled similarly to what dedicated marine-traffic tracking sites show, rather than plain static dots.
News Feeds
Most active forks bundle in 25 or more live 24/7 news broadcast streams, outlets like NBC, CBS, Sky News, Al Jazeera, France 24, and NHK are commonly named, playable directly by clicking a corresponding dot on the map. A number of forks also pull in public Telegram channel content as an additional real-time news signal, scraped from publicly accessible channels rather than private messages.
Is Osiris AI Free? Where to Get It on GitHub
Yes, Osiris is completely free, self-hosted dashboard software, released under an MIT license in most of its active forks. There’s no subscription, no paid tier, and no company behind it selling access, since it isn’t a hosted product at all, it’s a codebase you download and run yourself.
This is worth stating plainly because of how the searches around this tool tend to be phrased, including osiris ai github itself: there is no single official “Osiris AI” company or app to sign up with. What exists is a codebase, originally built as a Next.js dashboard by one developer and since forked and independently extended by several others on GitHub, each with a slightly different exact feature set, README, and setup process. Searching GitHub directly for “Osiris OSINT dashboard” will surface the currently active forks, and it’s worth checking a specific repository’s own README for its exact current feature list and setup steps before assuming every fork behaves identically, since they’ve diverged somewhat as different developers have extended the project in different directions.
How to Run Osiris AI Yourself
The core setup pattern is consistent across most active forks, since they share the same underlying Next.js and Docker structure, though you should confirm the exact repository URL and any recent changes against your chosen fork’s current README before running these commands.
What you need first: Docker installed on your computer (Docker Desktop on Windows or Mac, Docker Engine on Linux), and that’s genuinely it for most core features, since the project is built to work without requiring any API keys out of the box.
The basic setup sequence:
- Clone your chosen fork’s repository:
git clonefollowed by the specific repository URL from GitHub. - Move into the project folder:
cd osiris(or whatever the specific fork’s folder is named). - Copy the environment template file:
cp .env.template .env. This step is optional to edit further, since core feeds work without any keys added. - Start the application:
docker compose up -d. - Open your browser to
http://localhost:3000to see the live dashboard running on your own machine.
If you’d rather skip building the image yourself, several forks also publish a prebuilt image on GitHub Container Registry, letting you pull and run it directly with a docker pull command instead of a full clone and build. A couple of optional environment variables show up across forks too, adding OpenSky Network credentials or an N2YO API key for satellite tracking enhances certain data layers slightly, but the dashboard runs and displays its core layers without either.
If you’ve set up a self-hosted tool like this before, the pattern will feel familiar, it’s broadly the same Docker-and-localhost workflow covered in our guide to setting up SillyTavern, another self-hosted, open-source project you run on your own machine rather than through a hosted account.
How Does the AI Part Work?
Some of the more actively developed forks include a built-in AI “analyst” feature, a chat interface layered on top of the live map that can answer cross-domain questions grounded in whatever data the dashboard currently holds. The example the project’s own documentation gives is a genuinely representative one: asking something like “any military flights near active conflict zones?” and getting an answer that cross-references the live flight layer against the conflict zone layer in real time, rather than you manually eyeballing two separate map layers yourself.
This is a meaningfully different kind of AI feature from a general chatbot, since it’s not drawing on broad training knowledge to answer, it’s specifically querying the structured data Osiris already has loaded, flights, earthquakes, news events, whatever layers are active, and reasoning across them. It’s a similar underlying concept to Meta’s Muse agent connecting to your own apps to act on real data, covered in our Meta Muse AI agent guide, except here the AI is reasoning over public situational data rather than taking actions on your personal accounts.
Some of the more extended forks go further still, bundling in a genuine OSINT and reconnaissance toolkit alongside the map, DNS lookups, WHOIS queries, certificate inspection, and port or vulnerability scanning among the tools mentioned in project documentation. This is real security-research tooling, the same general category as tools like Shodan or Wireshark, and it’s exactly the part of the project where the Responsible Usage guidance below actually matters most.
Responsible Use Notes
Several active forks of this project include an explicit Responsible Usage policy, and it’s worth taking seriously rather than treating as boilerplate. The policy states plainly that the tools are intended for defensive, educational, and authorized monitoring purposes only, that you shouldn’t use any included scanning or reconnaissance tools against infrastructure or networks you don’t own or have explicit permission to test, and that compliance with applicable local, national, and international law is entirely the user’s own responsibility.
That framing matters because the honest reality of a tool like this is dual-purpose: the same recon toolkit that helps a security researcher audit their own systems could be pointed at something the user doesn’t have permission to touch, and the project’s own documentation is explicit that the second use is not what it’s built or intended for. Treat this the same way you’d treat any legitimate security tool, Wireshark, Nmap, Shodan, useful and completely legal for its intended defensive and educational purpose, and a genuinely different situation the moment it’s pointed at a target without authorization.
Common Questions
Is Osiris AI safe to run?
The core dashboard itself, built on well-established, widely used technology like Next.js and MapLibre, carries the same general risk profile as any self-hosted open-source project: reasonably safe when downloaded from an active, visible GitHub repository, and worth being more cautious about with any fork that looks abandoned, has very few stars or contributors, or asks for unusual permissions beyond what the core dashboard needs to function.
How “live” is the data, really?
This varies by layer rather than being one single answer. Flight positions and earthquake data update close to true real time, typically refreshing every few seconds to a minute, since both OpenSky and USGS are themselves live data feeds. News streams are genuinely live broadcast video. The conflict zone layer, by contrast, is closer to periodically refreshed than second-by-second, since it’s synthesized from news and event data rather than a dedicated live feed. If a specific layer seems to be lagging, that’s a difference in how directly each layer is sourced, not necessarily a bug in the fork you’re running.
Is it legal to use?
Using the dashboard itself is legal, since every data source it pulls from, flight tracking, public traffic cameras, earthquake data, public news, is already public and freely accessible on its own. The legal risk in this category of tool shows up specifically around the optional recon and scanning toolkit some forks include, running a port scan or vulnerability check against a system you don’t own or have explicit permission to test is illegal in most jurisdictions regardless of what software you use to do it, which is exactly the line the project’s own Responsible Usage policy draws.
Does it work on Windows, Mac, and Linux?
Yes, since the entire setup runs through Docker, which is cross-platform, the same setup steps work on Windows (via Docker Desktop), Mac (via Docker Desktop), and Linux (via Docker Engine) without needing separate platform-specific instructions.
How is this different from FlightRadar24 or a dedicated ship-tracking site?
Single-purpose trackers like a dedicated flight or ship tracker generally do one data type more polished and with a friendlier interface than Osiris does for that same layer alone. Osiris’s actual value isn’t out-competing any one of those individually, it’s combining several categories, flights, ships, earthquakes, news, cameras, that you’d otherwise need several separate tabs open to follow, into one map you self-host and fully control rather than relying on someone else’s hosted site staying online and ad-free.
Who actually built it, and can I trust it?
The project originated with one developer and has since been forked and independently extended by several others on GitHub, which is a normal and generally healthy pattern for an open-source project, more active forks means more eyes on the code, not fewer. That said, “trust” here is really a per-repository question rather than a blanket one: check a specific fork’s commit history, contributor activity, and issue tracker before running it, the same due diligence worth applying to any self-hosted open-source tool, and favor forks that are visibly active over ones that haven’t been updated in a long stretch.
Does Osiris AI cost anything to run?
No ongoing cost for the software itself, since it’s free and open-source. The only realistic cost is your own computer’s resources while it’s running, and optionally, if you choose to add API credentials for enhanced flight or satellite data, though the core dashboard functions fully without paying for any of that.
Is Osiris actually similar to Palantir?
Only in the loosest sense that both are software platforms built around visualizing and analyzing large amounts of data for situational awareness. Palantir is a large, well-resourced commercial company building software for government and enterprise clients with access to substantial private and often classified data. Osiris is a free, community-built dashboard pulling exclusively from public data sources, flight tracking, government traffic cameras, public earthquake feeds, open news. The comparison works as a rough category description, not a claim about matching capability or data access.
Do I need to know how to code to use it?
Basic use, once it’s running, doesn’t require coding, you’re clicking around a map interface in your browser. Getting it running in the first place does require comfort with a terminal and running a handful of copy-paste Docker commands, which is a real but fairly low bar for anyone who’s set up a self-hosted tool before.
Can I access it from my phone?
Since it runs as a local web server on whatever machine you install it on, you can reach it from your phone’s browser the same way you would any locally hosted web app, by visiting your computer’s local network address instead of localhost, provided your phone is on the same network. There’s no dedicated Osiris mobile app.
Why are there so many different versions on GitHub?
Because it’s genuinely open-source, any developer can copy (fork) the original project and continue developing it independently. Over time, that’s produced several actively maintained forks with somewhat different feature sets, some closer to the original map-only concept, others extended further into the cyber-intelligence and recon-toolkit direction. This is a normal pattern for popular open-source projects, not a sign of anything broken or inconsistent about the underlying idea.
Where to Find It
Osiris is a genuinely interesting example of how much real-time global situational data is already public and free to access, it just hasn’t historically been assembled into one browsable map before. Search GitHub directly for the current active forks, read the specific repository’s README for its exact feature list and setup steps, since they do vary, and if you’re comfortable with a basic Docker setup, you can have your own live global dashboard running locally within a few minutes.
For more explainers and setup guides like this one, check the blog, including our guides on setting up SillyTavern, Meta’s Muse AI agent, and AI marketing tools for small business if self-hosted and AI-powered tools are something you’re generally exploring. You can also visit the homepage for more guides.
Osiris is an actively developed, community-forked open-source project. Feature sets, setup commands, and repository URLs referenced here reflect the general pattern across active forks as of publication, always check your specific chosen repository’s current README before running setup commands, since individual forks update independently and on their own schedule.
