Tusi technical documentation
Guidance for building dependable AI systems
Technical guides, architecture patterns, and practical references for building and operating real-world AI systems.
$ request /api/health
{
"status": "healthy",
"checks": {
"database": {
"status": "up",
"latencyMs": 12
}
},
"version": "0.1.0"
}Recommended path
From a clear problem to a dependable system
This short path helps you define success and system boundaries before choosing technology.
- 01
Define the problem
Write down inputs, outputs, constraints, and a measurable success criterion.
- 02
Choose the right architecture
Decide between RAG, agents, computer vision, or a purpose-built combination.
- 03
Ship with observability
Design evaluation, security, monitoring, and rollback into the system from day one.
curl --request GET \
--url https://tusihub.ir/api/health \
--header 'Accept: application/json'Example response
{ "status": "healthy", "version": "0.1.0" }Learning paths
Find the topic you need
Guides are organized around the job you are trying to accomplish.
Response contracts, health checks, error handling, and secure API consumption patterns.
A path from problem definition and architecture choice to a production AI capability.
Type-safe integration examples for TypeScript, JavaScript, and Python.
Access controls, data protection, and threats specific to AI systems.
Documentation principles
Written for real engineering decisions
Every guide starts with trade-offs, system boundaries, and selection criteria.
Access controls, data protection, and failure modes stay in the main path.
Evaluation, telemetry, and runbooks are not an afterthought.
Still have a question?
Talk to a solutions engineer
Bring your current architecture, constraints, and business goal. We will help identify a sound technical path.