Persistent AI agents that never forget
Unlike disposable chat sessions, Fabricon agents are persistent entities with identity, memory, and specialisation. Each agent maintains long-term context across workflow runs — learning from past test results and improving over time.
Dual Memory Architecture
Agents maintain both long-term semantic memory across runs and short-term context within an active workflow, building knowledge over time.
Per-Agent LLM Selection
Assign a different model to each agent based on what it needs to do. For example, use a frontier model for the supervisor or test designer where reasoning matters most, and a faster, cheaper model for script generation or reporting — any combination that balances performance with cost.
Agent Facts
Each agent maintains a persistent fact store — domain knowledge, project context, team preferences — that accumulates and refines across every workflow execution.
Always-On Readiness
Agents remain warm and available 24/7. Trigger via schedule, webhook, or API call. No cold-start delays, no manual intervention required.
Agent Types
Test Engineer
UI Testing, API Testing, Integration Testing
QA Analyst
Test Design, Bug Reports, Quality Metrics
Security Tester
OWASP Top 10, CVE Scanning
Data Engineer
Test Data Generation, DB Validation
Env Engineer
Deployment, Health Checks, Data Reset
Test Lead
Review, Validate, Approve/Reject
Coordinator
Sprint Tracking, Stakeholder Reports
Extend agent capabilities with pluggable skills
Skills are instruction sets injected into agent prompts. Assign pre-built skills or create custom ones. Auto-assigned based on agent type, versioned, and shared across your organisation.
Test Automation
12 skillsSecurity
8 skillsReporting
6 skillsDevOps
8 skillsVisual workflow builder with zero-code orchestration
Deploy pre-built workflows in minutes or design fully custom pipelines to match your team's process. Dial in as much or as little autonomy as you need — insert human approval gates at any point, or let the platform run end-to-end on its own.
Node Types
Trigger
Cron schedule, JIRA webhook, or manual start.
Agent Task
Assign work to a specific agent type.
Supervisor Review
Test Lead validates output before proceeding.
Condition
Branch flow based on agent output or test results.
Parallel Split
Execute multiple agent tasks concurrently.
End
Workflow completion with optional notification.
Edge Conditions
Connect to any system via any protocol or standard
Fabricon connects to your entire application landscape — modern APIs, legacy enterprise systems, messaging platforms, and infrastructure — however your environment is built.
MCP Integrations
Live data injected into agent prompts
REST API Integration
Connect any API with full auth
More Ways to Connect
Additional protocols and integration patterns supported
Modern APIs
Legacy & Enterprise
Messaging & Events
Infrastructure
Browser-Based Testing
End-to-end UI and visual validation
Validate UI flows, visual state, and end-to-end user journeys across any enterprise front end — whether on-premises, cloud-hosted, or built on cloud-native services. Agents drive real browser sessions, capturing screenshots, asserting page state, and detecting regressions across every release.
Backend & Database Testing
API contracts, data integrity, service behaviour
Verify API contracts, query results, and data integrity across on-premises systems, cloud-hosted services, and cloud-native platforms alike. Agents test backend behaviour directly — independently of the UI — ensuring your infrastructure is reliable before changes reach production.
Ready to transform your test engineering?
Join engineering teams that ship faster with AI-powered test automation. Start your free trial today — no credit card required.