RQM Platform¶

RQM Technologies is a quaternionic, compiler-first quantum software stack built around SU(2)-native optimization, backend-agnostic public circuits, and workflow surfaces that extend from mathematical analysis to execution routing and Studio operations.

:material-lightning-bolt: API Quickstart :material-monitor: Studio Overview :material-book-open-variant: Architecture

What RQM is¶

RQM is not a generic SDK wrapper. It is a layered platform with clear technical boundaries:

Layer	Repository	Current role
Mathematical spine	`rqm-core`	Quaternion, spinor, SU(2), Bloch, validation, linear algebra, and coupling-analysis primitives
Optimization engine	`rqm-compiler`	Internal optimization and rewriting after the public circuit boundary; canonical 1Q IR is `u1q`
Service boundary	`rqm-api`	HTTP intake for validation, analysis, optimization, execution routing, auth, billing, chat, and TTS
Workflow layer	`rqm-studio`	Visual and operational layer for theory, optimization, execution-oriented flows, jobs, and account surfaces

The public circuit boundary remains rqm-circuits schema 0.2. Internal compiler IR is separate by design.

Where RQM fits in the stack¶

client / SDK / Studio
        │
        ▼
rqm-api  ← service boundary, policy, readiness, routing
        │
        ▼
rqm-circuits 0.2  ← public wire format boundary
        │
        ▼
rqm-compiler  ← internal optimization and rewriting engine
        │
        ├── canonical internal 1Q IR: u1q
        └── explicit later lowering for backend targets
        │
        ▼
backend adapters / execution providers

This is how RQM stays backend-agnostic at the public circuit boundary while still allowing explicit backend-targeted lowering and execution when requested.

What you can do now¶

Fetch valid named example circuits from the API (bell, ghz, optimizable)
Validate circuits against the current public schema and analyze them before optimization
Run optimization through a compiler pipeline that normalizes, canonicalizes, lowers to u1q, merges compatible 1Q structure, and cancels eligible 2Q structure
Inspect coupling / preservation signals through qualitative or measured analysis where supported
Check execution readiness with GET /v1/execute/capabilities before exposing backend choices
Route execution through Qiskit paths and Braket paths, including local simulation and held-job hardware flows where available
Use Studio as the visual and workflow layer for theory, optimization, execution-oriented tasks, jobs, reports, and Pro/account surfaces

Core platform boundaries¶

`rqm-core`¶

Owns the mathematical foundation. That includes quaternion and SU(2) primitives, Bloch/state utilities, validation helpers, linear algebra, and the coupling / entanglement analysis primitives used to support trust and preservation analysis.

`rqm-compiler`¶

Begins after the public rqm-circuits boundary. Named gates are lowered into the canonical internal single-qubit form u1q. Backend-targeted lowering is a later, explicit stage.

`rqm-api`¶

Exposes the current service surface: circuits, coupling analysis, execution, authentication, billing/wallet, and media-oriented endpoints. Responses follow a stable operational envelope with status, data or error, and meta.

`rqm-studio`¶

Sits on top of rqm-api. Studio is the product workflow layer, not the service boundary. It organizes user flows; the API remains canonical for service integration.

Why the compiler-first model matters¶

RQM optimization is proof-gated and fail-closed.

Only verified optimization candidates are committed
If verification cannot be established, the original circuit is returned unchanged
Backend-targeted lowering happens after optimization, not during public-circuit intake

That separation is central to the trust story: public interchange stays stable, optimization stays internal, and backend lowering stays explicit.

What is still evolving¶

RQM documents some surfaces conservatively on purpose:

Measured coupling analysis is not unrestricted; current measured coverage is limited and qualitative fallback remains part of the contract
Some optimization profile and backend-hint combinations are currently equivalent or reserved in the current release
Billing and Pro readiness do not by themselves guarantee hardware execution; provider readiness, credentials, validation, and quote/hold state still matter
Studio covers broader product workflows today, but route-level maturity is not uniform enough to describe every surface as equally deep

Start in the right place¶

Goal	Page
Make your first API call	API Quickstart
Understand the service surface	API Overview
Understand the visual/workflow layer	Studio Overview
Understand platform boundaries	Architecture
Understand optimization semantics	Optimization
Understand the internal canonical IR	Canonical IR (`u1q`)
Understand the broader stack	Ecosystem