STEP 01
Unified API
One base URL for text, image, and video
All projectsFIELD NOTE 04 / 2025
AI Infrastructure & Developer ToolsRAVAND.STUDIO / PROJECT 04
HMGateway — Unified AI Model Gateway
HMGateway is the developer platform behind HoosheMasnoei.com: a unified gateway that routes text, image, and video requests across hundreds of AI models and providers. Product and engineering teams can use a single API key and an OpenAI-compatible interface to integrate models quickly, monitor usage and cost in one dashboard, and stay resilient when a provider is unavailable. Technical documentation, model discovery and leaderboards, uptime monitoring, rial-based billing, and enterprise capabilities come together in one coherent developer experience.
Ravand roleStrategy, design, engineering, and release
Engagement1 Month
Accountable team3 people
DeliveredShipped
SPECIMEN / 04202504ROOT / BRANCH / OUTCOME
AI Infrastructure & Developer ToolsHMGateway — Unified AI Model Gateway
صدها مدل
Hundreds of models, one API
1 Month
Engagement
5
Technology & infrastructure
01
ROOT / THE PROBLEM
The problem
The challenge was not simply connecting a handful of models. Provider differences, cost, rate limits, and outages had to disappear behind a simple, predictable developer experience.
Research, product behavior, and operational data stayed connected through one shared account of the outcome.
CONSTRAINTS OBSERVED
- 01Move operational complexity behind one shared model
- 02Make impact observable from day one
- 03Protect user trust and clarity
02
BRANCH / THE RESPONSE
The response
Ravand created a single path from API key to production request: an OpenAI-compatible API, model routing, fallback, billing, and monitoring in one coherent product.
STEP 02
Routing and fallback
Move between models while keeping the service available
STEP 03
One dashboard
Manage keys, credits, usage, and cost in one place
03
OUTCOME / MEASURED
Measured outcome
The outcome is a production-ready product with a clear developer journey—from discovering a model and getting a key to sending requests, controlling cost, and monitoring reliability.
01
One API
02
Hundreds of models
03
Built-in fallback
04
PROCESS / 1 Month
Build process
The product was built in four deliberate phases so the experience and infrastructure became useful together: a shared contract, the gateway, operational controls, and the developer journey.
Shared contract
Define endpoints and OpenAI-compatible patterns
Gateway and models
Connect providers and the model catalog
Operational control
Credits, billing, monitoring, and fallback
Developer experience
Documentation, samples, and quick start
Principles
PRINCIPLES / 03Familiar standards
OpenAI-compatible tools reduce setup friction
Model choice
Switch models and providers without product lock-in
Operational clarity
Usage, credits, cost, and status in one view
Reliability by default
Fallback routing and service observability
01
MODULE / RESOURCES
Resources
The main paths for evaluation and onboarding—from the model catalog to the complete API reference and service status—are available in one place.
02
MODULE / STACK
Technology stack
The stack brings the web experience, gateway services, and integration tooling together in clearly separated layers.
Next.js
Platform UI, dashboard, and documentation journeys
NestJS
Backend services and gateway logic
OpenAI-compatible API
Unified contract for model integrations
LangChain
Integration path for LLM applications
Swagger
Interactive endpoint reference
DELIVERY LEDGER
Delivery ledger
The final delivery went beyond an API: it combined the operational surfaces and developer experiences needed to move from evaluation to production.
Multi-model gateway01
Developer dashboard02
Catalog and leaderboards03
Docs and monitoring04
NEXT SIGNAL / OPEN
The next problem can take root in one clear conversation.
If your problem sits between product, engineering, and growth, let us define it precisely together.
Start a conversation