Crisis Input orchestrates an entire simulated information environment — coordinated news coverage, social media storms, email floods, and live TV broadcasts — so your teams face the pressure of a real crisis, not a conference-room discussion.
Tabletop exercises gather your team in a room to talk through hypothetical scenarios. Participants discuss what they would do — but never face the cognitive overload, conflicting information, and time pressure of an actual crisis. The result is false confidence.
Inject-by-email tools improve slightly on the format, but require facilitators to spend weeks writing content, manually send stimuli, and track responses on spreadsheets. There is no real media environment, no social media dynamics, and no objective way to measure performance.
Meanwhile, regulators under DORA, NIS2, and SOX increasingly expect documented evidence that your teams can execute under pressure — not just that they attended a meeting.
The gap is clear: organizations need exercises that replicate the multi-channel chaos of a real crisis and produce measurable results. That is precisely what Crisis Input delivers.
Crisis Input is a full-stack crisis simulation platform that lets you design multi-channel exercise scenarios, launch them with automated orchestration, and measure your team's performance with objective data. Participants interact with simulated news sites, a social media platform, an email system, and live TV broadcasts — all running simultaneously, all controlled from a single scenario engine.
Build your scenario using pre-built inject packs, AI-generated content, or custom materials. Define phases, assign teams, set trigger rules, and configure which channels are active. Clone and adapt previous scenarios in minutes.
Start the exercise and let the scenario engine take over. Injects fire automatically based on time, participant actions, or inaction. Facilitators observe and intervene only when necessary — no more puppet-mastering every email and news article.
Automated KPIs capture detection time, reaction speed, decision quality, and task completion. Weighted evaluation grids let you score each team and phase. Replay the exercise second by second for debriefs. Export board-ready reports.
Participants face coordinated news articles, social media backlash, urgent emails, and live TV coverage at the same time — just like in a real crisis. When the pressure is real, people stop theorizing and start acting. That is where genuine preparedness begins.
AI content generation across four LLM providers produces realistic articles, social posts, emails, and stakeholder reactions in minutes. Pre-built inject packs cover major crisis types. Scenario cloning with deep copy means you never start from scratch.
Conditional trigger rules deliver injects based on elapsed time, participant inaction, decisions made, phase changes, or custom thresholds. The scenario adapts to your team's actions. Facilitators monitor the exercise — they don't operate it.
Automated KPIs measure detection time, reaction speed, task completion rate, and media pressure response. Weighted evaluation grids with 11 criterion categories replace subjective facilitator impressions with quantitative, board-ready metrics that you can compare across exercises.
A unified exercise log captures every event across all channels with microsecond precision. Scrub through the timeline to see exactly who did what, when, and how the situation evolved. Debriefs become precise, evidence-based discussions — not memory contests.
Self-hosted via Docker with no cloud dependency. Run AI content generation fully offline using Ollama. Your scenario data, participant information, and exercise results never leave your infrastructure. Full data sovereignty, guaranteed.
Crisis Input runs five distinct communication channels simultaneously, each with the depth and fidelity of a standalone product. Simulated news outlets publish articles with different editorial tones, credibility ratings, and scheduling — from breaking news banners to background noise. Squawk, the built-in social media platform, reproduces follow graphs, trending algorithms, bot amplification, and viral dynamics. The internal email system supports threading, distribution groups, attachments, priority levels, and realistic delivery failures. A corporate website hosts org charts, document libraries, and press releases. Each channel operates independently but feeds into a single unified timeline.
Define multi-phase exercises where each phase sets its own media pressure level and simulation speed. Injects are delivered automatically through six trigger types: time-elapsed, inaction detection, decision-based, phase-change, threshold-based, and manual override. If a team fails to issue a press statement within thirty minutes, the social media pressure escalates. If they make a specific decision, a stakeholder reaction fires. Scenarios respond to participants — they are dynamic, not scripted. Pre-built inject packs by crisis type (cyber, HR, ESG, fraud, industrial) give you a head start on any scenario.
Generate hundreds of realistic injects using four LLM providers: OpenAI, Anthropic Claude, Mistral, or fully local Ollama for air-gapped environments. Prompt templates cover articles, social posts, emails, stakeholder reactions, and press releases — each calibrated for tone, format, and crisis context. Every AI-generated piece requires human approval before it enters the scenario. Token usage tracking and cost estimation keep budgets under control. No vendor lock-in: switch providers or run everything locally.
No other crisis simulation platform offers this. Upload base broadcast footage and schedule crisis-related video clips to insert at precise moments during the exercise. The HLS video stream delivers the experience of watching live television news as events unfold. Breaking news banners, reporter segments, and expert commentary appear exactly when the scenario demands them. Participants experience the same media pressure that defines real-world crises — the constant drumbeat of live television coverage.
Weighted evaluation grids cover 11 criterion categories: detection, initial reaction time, decision quality, internal coordination, external communication, escalation management, stakeholder handling, media response, recovery actions, documentation, and post-crisis learning. Each criterion is scored and weighted to reflect your organization's priorities. Automated KPIs — detection time, reaction speed, task completion rate, media pressure score — are calculated in real time. Phase-level analysis reveals where performance dropped. Self and peer evaluations add qualitative depth. Export everything to CSV for your compliance files.
Every action across every channel is logged in a single timeline with microsecond precision. Replay mode lets you scrub through the entire exercise chronologically: see which articles published, which emails were sent and read, how social media narratives evolved, and what decisions were made — all synchronized. The full audit trail records who did what, when, with old and new values for every change. Export to JSON, CSV, or PDF for compliance evidence supporting DORA, NIS2, and SOX requirements.
Data breach, ransomware, supply-chain compromise
A ransomware attack hits your infrastructure. Internal emails alert IT teams while news outlets report a data breach based on leaked screenshots. Social media accounts amplify the story with speculation about compromised customer records. The crisis cell must coordinate containment, legal notification, regulator communication, and public messaging — all while media pressure escalates on a conditional timer. Crisis Input's inaction triggers ensure that delayed responses produce realistic consequences.
Explosion, contamination, workplace fatality
An explosion at a production facility triggers a cascade of events. Live TV broadcasts show aerial footage of the site. News outlets publish conflicting casualty reports. Local residents flood social media with photos and demands for answers. Internal emails deliver fragmented field reports. The operations team must manage the physical response while PR handles media inquiries and the legal team assesses liability — simultaneously, under time pressure, with incomplete information.
Executive scandal, product recall, whistleblower leak
An investigative journalist publishes allegations about a senior executive. The story spreads across social media within minutes, amplified by bot accounts and trending algorithms. Employees share internal concerns via email. Stakeholders — investors, board members, regulators — demand responses. The team must decide on messaging, spokesperson strategy, and internal communication while the narrative evolves in real time. Decision-based triggers shape the scenario: every public statement changes what happens next.
DORA, NIS2, SOX exercise documentation
DORA requires financial entities to conduct threat-led penetration testing and resilience exercises. NIS2 mandates incident response preparedness across essential services. Crisis Input produces the evidence: timestamped logs of every action, weighted evaluation scores, automated KPIs, and exportable reports in CSV, JSON, and PDF. The full audit trail documents who participated, what decisions were made, and how the team performed — exactly what auditors and regulators need.
| Capability | Tabletop exercises | Inject-by-email tools | Generic LMS | Crisis Input |
|---|---|---|---|---|
| Realism | Low — verbal discussion only | Moderate — email channel only | Low — content delivery, no simulation | High — five simultaneous channels including live TV |
| Automation | None — fully manual facilitation | Basic — scheduled email sends | Course delivery only | Full — conditional triggers adapt to participant behavior |
| Scalability | One room, one session | Limited by facilitator capacity | Scales for content, not simulation | Unlimited concurrent exercises with scenario cloning |
| Measurement | Subjective facilitator notes | Manual response tracking | Quiz scores only | Automated KPIs, weighted scoring grids, phase analysis |
| Media simulation | Verbal description of media scenario | None | None | News, social media, email, live TV, corporate website |
| Time to deploy | Weeks of manual preparation | Days of content writing | Not applicable | Hours — AI generation, inject packs, scenario cloning |
Crisis Input is built for organizations where data sovereignty, auditability, and infrastructure control are non-negotiable.
Docker-based deployment with no SaaS dependency. Your data stays on your infrastructure. Air-gapped operation fully supported.
Every action logged with actor, timestamp, and change details. Export to JSON, CSV, or PDF. Supports compliance evidence for DORA, NIS2, and SOX.
CSRF protection, HSTS, Content Security Policy, X-Frame-Options, optional two-factor authentication, and IP whitelisting per scenario.
REST API with OpenAPI/Swagger documentation. Integrate with your SIEM, ITSM, or custom tools. Every platform capability is API-accessible.
Choose from four AI providers — OpenAI, Anthropic Claude, Mistral, or Ollama for fully local operation. Switch providers at any time without data migration.
Built on Django, PostgreSQL, Redis, and Celery — battle-proven technologies trusted by enterprises worldwide. Multi-language support: English and French.
Tabletop exercises test discussion. Crisis Input tests execution under multi-channel pressure — which is what an actual crisis demands. Participants face simultaneous news coverage, social media escalation, email floods, and stakeholder pressure. They must make real decisions under time pressure, not describe what they would hypothetically do. The gap between "we discussed it" and "we performed under pressure" is exactly what separates prepared organizations from unprepared ones.
You don't need to build them from scratch. Pre-built inject packs cover major crisis types — cyber, industrial, reputational, HR, ESG, and fraud. AI content generation produces realistic articles, social posts, emails, and stakeholder reactions in minutes across four LLM providers. Scenario cloning with deep copy lets you adapt a previous exercise for a new team or sector in hours, not weeks. Most organizations reduce preparation time by 80% compared to traditional methods.
Crisis Input produces quantitative, comparable metrics. Automated KPIs measure detection time, initial reaction speed, task completion rate, and media pressure response. Weighted evaluation grids score each team across 11 criterion categories. Phase-level analysis shows exactly where performance improved or degraded. These are board-ready numbers — not subjective facilitator impressions. Compare results across exercises to demonstrate measurable improvement over time.
Crisis Input is fully self-hosted via Docker. There is no SaaS dependency and no data leaves your infrastructure. For AI content generation, you can use Ollama to run language models entirely on your own hardware — no external API calls required. The platform supports air-gapped environments. You control every aspect of the deployment.
Participants interact with interfaces they already know: a news website, an email inbox, a social media feed, a TV broadcast. There is no training required. The platform is designed so that participants focus on the crisis, not the tool. The complexity lives in the scenario engine and orchestration layer — which only facilitators and administrators need to understand.
Training platforms deliver content. Crisis Input orchestrates an environment. Participants don't learn by reading slides or watching videos — they learn by operating under simulated pressure across multiple channels. It is the difference between reading about swimming and being placed in water. Crisis Input is an exercise platform, not a course platform.
Crisis Input deploys via Docker Compose. The stack includes Django, PostgreSQL, Redis, and Celery with WebSocket support for real-time updates. Minimum requirements are modest: a modern Linux server with Docker installed. For AI content generation with Ollama, a GPU is recommended but not required. The platform ships with development and production Docker Compose configurations, ready to deploy.
Yes. Crisis Input exposes a full REST API documented with OpenAPI/Swagger. You can integrate with SIEM platforms, ITSM tools, identity providers, or custom internal systems. Every capability available in the interface is also available via API.
Crisis Input supports team structures with over ten role categories — Crisis Cell, Legal, PR/Communications, Operations, Spokesperson, IT, HR, Executive Committee, and more. Each participant receives their own accounts across all active channels. There is no hard limit on participants per exercise. The platform supports multiple concurrent exercises with separate scenarios and team structures.
Crisis Input supports English and French. All interface elements, evaluation criteria, and export templates are available in both languages. Scenario content — articles, posts, emails — can be written or AI-generated in any language supported by the chosen LLM provider.
Regulatory deadlines under DORA and NIS2 are approaching. Board expectations for documented crisis preparedness are rising. And the next incident — cyber attack, industrial failure, reputational shock — will not schedule itself around your readiness. See how Crisis Input can transform your crisis exercises from checkbox activities into genuine operational tests.
Not ready to talk yet? Download the product brief for a complete technical overview.