Model. Generate. Run.
Netgrif is the AI-Native Process Platform powered by Petriflow — a rigorous, graphical, human-readable language built for AI-driven automation. One model captures process, data, UI, and roles in a single artifact. The platform interprets it directly — no compile step, no translation layer between specification and running application.
PLATFORM
Powered by Petriflow
An AI-native modelling language designed for LLMs and humans alike. Deterministic enough to trust, expressive enough to build with — one model unifies process, data, UI, and roles. Specification equals application.
PLATFORM
Built on a modern open-source stack
Java/Spring Boot, Angular, MongoDB, Elasticsearch — trusted, scalable, and transparent. Netgrif runs on your infrastructure or in the cloud, with no vendor lock-in.
Key Benefits
One Model, Whole Application
Process, data, UI, and roles live in a single Petriflow artifact. The LLM sees the whole app in one context window — no fragmenting across layers, no inconsistency between code, schema, and frontend. Internal eTasks and external eForms run from the same model, with synchronized logic end-to-end.
Specification == Application
Petriflow models aren’t diagrams that get translated into code — they are the code. The platform interprets them directly. No compile step, no scaffolding, no synchronization between visual design and implementation. Edit the model, the running app changes.
From AI prompt to rich code
Describe the workflow in natural language and the AI builder generates executable Petriflow. Refine visually in the drag-and-drop builder, or drop into rich-code when you need full control — without losing the model as your source of truth.
One Environment, Endless Use Cases
Replace fragmented tools, shadow IT, and disconnected workflows with one consistent platform. HR, procurement, compliance, customer services, finance — every agenda runs as a Petriflow application on the same runtime. Same language, same engine, same governance.
Validate Petriflow models before deployment through simulation — confirm logic, role permissions, and data flow behave as intended. Once live, monitor state transitions and task execution in real time to detect bottlenecks, optimize performance, and gain immediate operational insight.
Modify and redeploy Petriflow models instantly — no service restart, no impact on running cases. Adapt to change, ship fixes, or roll out improvements while the platform stays up. The interpreter picks up the new model without breaking in-flight work.
Petriflow bridges business analysts, developers, and AI. The same model used in initial design carries through deployment, iteration, and AI generation. No translation gap between BPMN diagrams, code, and runtime — the model is the single source of truth across the entire SDLC.
Built ground-up for AI generation, not retrofitted. The AI builder turns natural-language requirements into executable Petriflow — 40–60× more efficiently than generating JavaScript or fullstack code. The model’s open architecture also integrates external AI services (NLP, document processing, prediction APIs) directly into workflows.
Built for End-to-End Automation
One AI-native platform, one language, every enterprise workflow:
- Replace fragmented legacy systems with one Petriflow runtime
- Generate enterprise applications from natural-language prompts
- Digitize manual tasks and orchestrate team collaboration end-to-end
- Serve internal users via eTasks and external customers via eForms — same model
- Stay compliant with built-in role-based access control (RBAC)
- Integrate with your IT ecosystem via SOAP and REST APIs