M13 SYSTEM LAYER
The system behind every model.
The m-pathy API does not return raw output, it returns controlled operations through M13.
WHY M13
See the operation behind the answer.
A normal model call gives you text, but M13 gives you the system layer around it.
You see:
what was executed
which model worked
which operation was used
what it cost
what was created
how it can be traced
So you can use AI output inside real workflows, not just read it.
RICH RESPONSE
M13 returns more than text.
An M13 response is a structured system object with generated content and runtime facts.
01
Content
The model output shaped by the active M13 operation.
02
Usage
Token usage, cached tokens and billable tokens.
03
Billing
Debit status, debit amount, balance source and balance after the call.
04
Artifact
A structured result object that can be stored, rendered or processed.
05
Server IRSS
Server side runtime evidence produced outside the model response.
06
Execution context
Run ID, stage ID, operation ID, command and execution metadata.
07
Ledger entry
A traceable reference for later audit and proof.
You do not just receive an answer.
You receive a controlled operation result.
FOUR API MODEL SPACES
Use API models by operation, not by guesswork.
M13 places each API model into a specific Space, so the operation defines what the model is allowed to do in that call.
Claude Sonnet 4-6M13 reasoning
Space 01
When you need structured thinking.
Use M13 reasoning when complex input needs to be processed into a stable analytical foundation.
M13 places Claude Sonnet 4-6 into a reasoning space for structured analysis, stepwise derivation and robust information processing.
Analysis
Planning
Structured derivation
Complex input processing
Workflow preparation
Claude Opus 4-6M13 challenge
Space 02
When you need to test before you trust.
Use M13 challenge when an answer, plan, report or decision should not move forward without a stress test.
M13 places Claude Opus 4-6 into a validation space.
The operation tests assumptions, missing evidence, contradictions, weak boundaries, unsupported inference, risk and drift.
C6 challenge
Validity checks
Contradiction testing
Risk reduction
Decision stress tests
Structural quality control
You do not receive a second opinion.
You receive a controlled challenge operation.
GPT-4.1 miniM13 summary
Space 03
When you need controlled compression.
Use M13 summary when long content needs to become clear, compact and usable.
M13 places GPT-4.1 mini into a summary space for controlled reduction, ordered output and cost-aware summarization.
Report summaries
Status output
Log compression
Short result forms
Cost-efficient processing
Claude HaikuM13 fast
Space 04
When you need fast orientation.
Use M13 fast when you need quick classification, first orientation or lightweight processing with low latency.
M13 places Claude Haiku into a fast operation space for simple checks, routing decisions and rapid pre-processing.
Quick checks
Light classification
First orientation
Routing decisions
Low latency workflows
API OPERATIONS
The m-pathy API does not forward prompts to models.
Many APIs give you model access, but m-pathy gives you controlled M13 operations with command, rich response and runtime evidence.
This helps you build AI features that are not loose text surfaces.
You can build workflows where model output is structured, billable, traceable and ready for downstream processing.
WHAT YOU CAN BUILD
Use M13 when AI output must become part of a system.
Challenge a plan before your workflow accepts it.
Summarize long results into a compact response object.
Separate reasoning, summary, challenge and fast operations.
Track usage and billing for every operation.
Store artifacts for later rendering or processing.
Connect API results to ledger evidence.
Build extensions that use M13 operations.
Choose models by system function, not by intuition.
DEVELOPER PATH
Start from the developer path.
You start with the developer path, not with a loose API key.
The Build page guides API access, Developer Schooling, SDK workflow and staging. Developer Schooling then opens the M13 Extension SDK workspace for local VS Code development.
M13 provides the infrastructure.
You own the runtime.
Your API key comes from M13.
Your private key stays local.
Your public key can be exported.
Your extension logic runs in your workspace.
BUILD AGAINST M13
Do not build against raw model output.
Build against M13 when AI should not only generate text, but operate inside a controlled system layer that m-pathy turns into an executable ecosystem.