Architecture
One application model and one runtime can run many different applications.
Applications are definitions. One runtime interprets them all, so every app gets the same access rules, audit and automation.
Metadata store
- entity Invoice
- screen Invoice list
- lifecycle Draft → Paid
- role Finance: Team
- workflow On approval
Applications
- Field serviceDevelopment · QA · Production
- ProcurementDevelopment · QA · Production
- Customer portalDevelopment · QA · Production
What is inside
The layers
Front ends on top, .NET services and engines in the middle, standard infrastructure below.
- Runtime web app with Form.io and 33 components
- Custom front ends on the client SDK
- TAF Studio
- TAFI in Studio and Claude Code
- Identity
- Metadata
- Data
- Forms
- Workflow engine
- Scheduler
- Notifications
- Reports
- Files
- Templates
- Code executor
- Logging and audit
- Query engine
- Hook pipeline
- Permission engine
- Lifecycle engine
- Approval engine
- Trigger engine
- PostgreSQL
- SQL Server
- RabbitMQ
- Redis
- Elasticsearch
- Docker
How it works
Every save takes the same path
Ordered checks before the write, one event after it. Platform features and your C# plug into the same pipeline.
- Accessrole scopes
- Planentitlements, seats
- Approval lockrecords under approval
- Ownershipworkspace rules
- Auditrecord info
- Your C#validate, pre, post
- Triggers
- Workflows
- Notifications
- Audit trail
- Global search
- Live grid refresh
Your data
Your database, per environment
PostgreSQL or SQL Server, one database per application environment.
Fast and standard
Cached where it pays, shipped the standard way
- Metadata cacheIn memory and Redis, per app and tenant
- Access rules cacheEvent-driven invalidation
- Compiled scriptsCached by content
- App schemaKept in the browser by the runtime
- Docker imagesServices ship as Docker images
- OpenAPIServices are documented with OpenAPI
- .NET SDKsMost services ship one
- API referenceThe services and their endpoints
Proof
The platform in numbers
For technical evaluators: the size of what sits underneath your application.
Counts are based on the current product codebase and may change between releases.
Limits
What this architecture does not guarantee
A reusable foundation does not by itself set your capacity, uptime or recovery targets. Here is what we do not claim.
- No published capacity figureWe do not publish a user, throughput or uptime number. Load test your own workload before you rely on one.
- Backups and disaster recoveryOn the roadmap. No recovery targets are published yet.
- Health checks on every serviceStandard health endpoints on each service are on the roadmap.
- Several copies of each serviceRunning more than one copy of a service, to handle more load, is on the roadmap.
See TechAppForce on your own use case
Explore each part of the architecture in detail. Select each component in the architecture guide, or walk through it with the engineers who build it.