There is a particular kind of uncertainty that settles into a society when violence becomes familiar. They live it, in unlit streets, overcrowded taxis, in unresolved grievances and random moments that turn into disaster then a statistic. Incremented by systems that only react once harm has already been done.
Across the continent, public safety is still structured around aftermath. A report after the crime, a patrol after the damage, an investigation after the loss. Between what is happening now and what is to come lies a gap filled with statistics; of complaints, of victims and of death. So we filled it.
With Salama.
the solution.
There is a kind of power in knowing what is coming.
In understanding that when systems can see patterns, understand pressure, and learn from behaviour, they can protect without panic and respond without delay. Salama was built around the belief that safety should be proactive, shaped by foresight, not aftermath. It was designed from actual experiences that shape everyday lives: from how people move, where risk concentrates, how tension escalates, and how violence repeats when signals are ignored. It transforms safety from a reactive function into a continuously learning, predictive, and self-optimising organism that connects the street to the station, the border to the city, the emergency call to the response, and the incident to its wider pattern.
Salama was made to see through the blind spots; seeing society, modelling instability, protecting infrastructure, securing borders, coordinating emergencies, enforcing accountability, and safeguarding vulnerable populations. In practice, it turns every patrol, every camera, every report, every movement, every interaction, and every decision into part of a single safety intelligence pool. By turning scattered signals into shared intelligence, we respond to trajectories, not incidents. We deploy to behaviours, not locations and investigate networks not cases, allowing institutions to intervene earlier and anticipate rather than chase.
our answer.
Our work begins before the incident, before the call, before the damage, in the quiet signals that other systems are not built to notice.
Salama is a sovereign public safety intelligence layer that unifies contextual awareness, behavioural modelling, predictive deployment, and institutional memory into a single national reasoning organism. Its agents run on an open-weight model post-trained on past incidents, replayed as they stood on the day with what happened next hidden. It is penalised for pointing at the wrong person, because here a false alarm is a harm in itself. The agents fuse visual, thermal, and acoustic sensing to detect concealed weapons, abnormal heat signatures, gunshots, crowd agitation, panic acoustics, and movement in unlit spaces. They read posture, gait, density, dispersal, and interaction to understand crowd pressure before unrest, and interpret vehicles as behaviour, linking reckless driving, route repetition, convoy formation, freight deviation, and cross-border movement to organised crime, trafficking, smuggling, and corridor activity. Physical signals are continuously aligned with social and digital signals, such as protest language, mobilisation cues, misinformation, recruitment, and agitation, creating a layered early-warning system where escalation is identified while it is still narrative, not violence. Digital crime and online harm are fully integrated into public safety. Cyber incidents, fraud, extortion, trafficking recruitment, and identity abuse feed into the same intelligence fabric, linking digital evidence and financial patterns with physical activity. Together, these signals power a living national safety model that connects economic pressure, service disruption, climate stress, and political events to crime, unrest, and migration, allowing leaders to simulate decisions, anticipate escalation, and govern with foresight rather than reaction.
Within this intelligence environment, policing itself becomes software. Agents design patrol geometry dynamically, automate dispatch, orchestrate emergencies and disasters, and simulate public-order events before they occur. Investigations operate as connected graphs not case files, allowing cases to accumulate memory; evidence and forensics are tamper-evident and linked across scenes; custody, courts, and outcomes close the loop from street to judgement. Agents protect infrastructure, energy assets, transport networks, borders, airspace, and corridors through continuous anomaly detection and aerial intelligence. Institutional integrity is enforced structurally, with misconduct, corruption, collusion, and procedural manipulation detected early. Womaner operates as a dedicated protection engine for women, ensuring that patterns of abuse, harassment, and exploitation are identified early, linked correctly, and acted on decisively. Crime reports, tips, voice notes, location sharing, and danger signals are fused with top-down sensing, allowing the system to learn which signals matter and which patterns indicate real risk. Child protection and missing persons are handled as intelligence cases, enabling rapid alerts, pattern linking, grooming detection, school safety protocols, and coordinated recovery.
An officer makes every decision that affects a person's liberty.
Salama is built for safety.
