CarbonLayer Demo Kit · Model-Based Simulation for Validation

Model-based simulation for validating inference cost, latency, and impact-evidence needs.

Explore inference cost and latency trade-offs, then identify the impact evidence your team needs to evaluate a workload.

Use the written walkthrough to scope fit. Any paid delivery is separate and requires agreed scope and terms; no pilot price, service value, or delivery date is offered here.

Sample receipt · 880 tokens
A prototype response shape; impact data is unavailable and latency is illustrative.

Simulation

simulated

Provider

not called

Impact data

unavailable

Latency · seeded

78 ms

Sample output is simulated; this page does not make an inference call.

Simulation disclosure

What this prototype does and does not measure

The DemoKit and /api/inference use simulated model output and route metadata; no model provider is called by that API. Impact data is unavailable. Region targets and latency are illustrative, not physical routing or production tracking. Open the separate cost playground to inspect provider cost and latency trade-offs; a configured provider call there is a direct call, not CarbonLayer live routing. Per-inference environmental values, energy allocation, baselines, and savings remain unavailable. The researcher-reported water summary is not independently verified; the full ledger and source PDF are absent. No candidate water values or confirmed water-impact claims are published.

Beat by beat · 3:00 total

The three-minute tour, written out.

Every receipt in this walkthrough is visibly bounded: output and routing are simulated, no model provider is called, and impact data is unavailable.

0:00

The gap: request logs need clear boundaries.

Explore inference cost, latency, and impact-evidence needs with clearly labeled simulated output.

Screen 1 · the homepage

0:25

The request: one familiar REST call.

Send a simulated model label, prompt, token budget, and latency preference. No model provider is called.

Screen 2 · the docs quickstart

0:55

The receipt: route metadata, no impact claim.

The response echoes the prompt and records the simulated route target, latency, and unavailable-impact status.

Screen 3 · the response contract

1:35

The ledger: a key-scoped request log.

GET /api/savings is authenticated and scoped to the credential. It returns receipt status, not environmental totals.

Screen 4 · the scoped receipt log

2:05

The network: seeded route targets.

GET /api/network is public and returns illustrative locations and route windows. Numeric environmental values stay out of the public contract.

Screen 5 · the public route snapshot

2:35

Next: scope a representative workload.

A fit-and-scope discussion can identify acceptance criteria. Any delivery is separate and requires agreed scope and terms.

Screen 6 · the pricing snapshot

No-provider disclosure

The /api/inference response is real, but its model output and route selection are simulated. Impact data is withheld rather than presented as zero or an estimate. See the methodology release record.

Annotated screens

The pages behind the walkthrough.

Open the current product, API, and plan pages for their live disclosures.

Home
Simulation and validation focus
API reference
Simulated receipt and status fields
Pricing
Availability-qualified plan previews

Representative API shape · sample content

Receipts with their limits stated.

The request and response below are representative sample content, not a live request, captured production receipt, or proof of production behavior. Model and route metadata are simulated and illustrative; per-inference environmental values, energy allocation, baselines, savings, SCI, and cost claims remain unavailable.

Request · POST /api/inference
Replace the visibly sample key after signup.
bash
curl -sS -X POST https://carbonlayer.polsia.io/api/inference \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"prompt":"Draft a status update for a city sustainability team.","model":"claude-sonnet","tokens":880,"preferLowLatency":false}'
Receipt · sample JSON
Representative sample JSON; no model provider call and no impact value.
json
{
  "model": "claude-sonnet",
  "tokens": 880,
  "response": "Draft a status update for a city sustainability team.",
  "device": { "name": "Oslo Harbour Hydro Cluster", "region": "eu-north-1", "latencyMs": 78 },
  "simulationStatus": "simulated",
  "modelProviderCalled": false,
  "impactDataStatus": "unavailable"
}

simulated

Simulation status

not called

Model provider

unavailable

Impact data

78 ms

Illustrative latency

Marseille status update
claude-sonnet · 880 tokens
StatussimulatedImpactunavailableLatency78 ms
EU route summary
claude-haiku · 120 tokens
StatussimulatedImpactunavailableLatency78 ms
Support-ticket classifier
cl-compact-v1 · 60 tokens
StatussimulatedImpactunavailableLatency78 ms

The researcher-reported water summary is not independently verified. The full ledger and source PDF are absent, so no candidate water values or confirmed water-impact claims are published.

GET /api/network · public simulated snapshot

Seven route targets, no impact claim.

The public endpoint returns region names, coordinates, and latency metadata for illustrative routing. Numeric environmental values are intentionally absent.

Open the current /api/network JSON snapshot

Oslo Harbour Hydro Cluster
eu-north-1 · Norway
Latency78 msImpact dataunavailable
Quebec Hydro Edge
us-east-1 · Canada
Latency92 msImpact dataunavailable
São Paulo Hydro Roof
sa-east-1 · Brazil
Latency184 msImpact dataunavailable
Tasmania Wind + Hydro
oc-east-1 · Australia
Latency220 msImpact dataunavailable
Columbia Gorge Wind Pod
us-west-2 · United States
Latency110 msImpact dataunavailable
Andalucía Solar Lab
eu-west-2 · Spain
Latency150 msImpact dataunavailable
Karnataka Renewable Bridge
ap-south-1 · India
Latency168 msImpact dataunavailable

Founder’s Circle / Beta Access Plans

Availability-qualified previews.

The existing site-listed prices and monthly inference allowances are preserved here and on /pricing; this update does not independently verify current commercial terms. Inference counts are usage allowances, not evaluation or delivery services. Scheduler execution, live Routing, and CSRD reporting are planned.

Free
Prototype access with simulated receipts; availability-qualified preview.
$0/ month

50,000 inferences / month

Simulation metadata only; paid availability depends on configuration.

Pro
Availability-qualified preview for teams evaluating the prototype API.
$29/ month

5M inferences / month

Simulation metadata only; paid availability depends on configuration.

Team
Availability-qualified preview for higher-volume prototype workloads.
$99/ month

25M inferences / month

Simulation metadata only; paid availability depends on configuration.

CSRD is a roadmap goal: ESRS mapping is currently in the mapping phase. No CSRD compliance, compliant report, or reporting capability is available through these previews. Confirm access and terms before purchase.

Priority-based roadmap

Now / Next / Later

Sequence reflects current priorities and carries no delivery date or guarantee.

Published · available
Methodology v1.0 evidence framework
The published framework is available on /methodology.
Queued · pending
Final coefficient verification
Coefficients are not verified. Per-inference carbon and water impact outputs remain unavailable.
Now
Label public impact values by evidence status.

Planned

Label public Playground impact values as measured, modeled, estimated, or unavailable. The labels preserve the evidence boundary; per-inference environmental values remain unavailable where supporting evidence is missing.
Next
Downloadable per-inference decision receipt.

Planned

A downloadable per-inference decision receipt containing route, latency, cost, impact values, confidence labels, and methodology version. This is a roadmap item, not a current capability or guarantee; unsupported impact values remain unavailable.
Later
Production capabilities remain planned and unverified.

Scheduler execution

Planned · not currently available or verified

Live provider or workload routing

Planned · not currently available or verified

CSRD reporting

Planned · not currently available or verified

ESRS mapping is in the mapping phase; it is not compliance or reporting.

On hold

Part 5

On hold

Coordinated campaign

On hold
Developer Console
Conditional

Work remains conditional until the survey is published and response counts are known. No demand conclusion is drawn.

Evaluation and delivery

Scope fit before discussing delivery.

Share workload details through the contact form for a fit-and-scope discussion. That evaluation does not promise delivery. Any paid delivery is separate and requires agreed scope and terms; no pilot price or service value is published here. Phase 1 will not be described as manually deliverable until its acceptance criteria are demonstrated end-to-end on a representative workload.

Discuss fit and scope
Four lines to include
These details help define evaluation scope.
  1. 1Model labelA rough description supports scope discussion only; it is not delivery acceptance.
  2. 2Approximate request volumeA rough description supports scope discussion only; it is not delivery acceptance.
  3. 3Regions where latency mattersA rough description supports scope discussion only; it is not delivery acceptance.
  4. 4Workload timingA rough description supports scope discussion only; it is not delivery acceptance.