BRICQS is an AI cloud platform that gives developers everything needed to build and run AI products in production: GPU deployments, managed databases, edge routing, domains, storage, and monitoring — all wired together and accessible through a unified API, dashboard, and CLI.

Every resource on BRICQS is backed by real infrastructure. When a deployment is running, it is a real Azure Container App. When DNS propagates, it is a real DNS zone. The platform surface is thin by design; the infrastructure it manages is real.

The five runtimes

LLM API

Connect OpenAI, Anthropic, or Google — unified OpenAI-compatible endpoint, secrets vault-backed.

POST /runtime/{project_id}/chat/completions

GPU Deployments

Deploy LLaMA, Mistral, Phi-3, Whisper on dedicated T4 GPU containers. Three environments with one-click promotion.

POST /deployments/deploy

Agents

Tool-using agent loop (Anthropic-backed) with calculator and HTTP tools. Up to 6 reasoning steps with full trace output.

POST /agent/run

RAG

Upload PDF/DOCX/TXT — BRICQS chunks, embeds, and answers questions grounded in retrieved content.

POST /rag-project/query

Workflows

Multi-step AI pipelines. Each step can call any runtime — LLM, RAG, Agent, GPU deployment — with template variable passing.

PUT /workflow/{project_id}/definition

AI Product Lifecycle

Every product on BRICQS moves through a 9-stage lifecycle, automatically tracked in the Product Registry.

01Plan02Build03Configure04Test05Evaluate06Release07Deploy08Monitor09Improve