Roofline website screenshot

Roofline

Roofline (RooflineAI GmbH, Köln, Germany) is a 2024 RWTH Aachen University spin-off building a retargetable AI compiler stack for edge deployment. Its three products are an SDK that compiles models ahead of time from any framework — the product page shows a single Python call, attic.compile(model, (input,)) — a lightweight C runtime with Python bindings that executes the compiled model and orchestrates heterogeneous SoCs, and a Performance Dashboard fed by a roofbench command-line tool running nightly benchmarks across hundreds of models for latency, throughput and memory footprint. The stack is built on open-source MLIR and IREE with proprietary hardware-specific optimization on top, and targets NPUs, MCUs and heterogeneous SoCs with quantization (INT8, blocked, symmetric, asymmetric, dynamic), dynamic shapes for on-device LLMs and mixture-of-experts support. Published work includes NXP eIQ Neutron NPU LLM enablement, Vulkan 1.3 and Arm SVE support in IREE, and day-0 support for Liquid AI LFM2. Roofline is a strategic partner of the Edge AI Foundation and raised a seed round from Serena and First Momentum Ventures. It sells compiler licenses to hardware vendors and embedded software teams: there is no developer program, no public API, no documentation host and no registry-published package. Every product page ends at a Contact Sales form, and the only machine-readable document Roofline publishes anywhere is an llms.txt on its marketing site.

Roofline is profiled on the APIs.io network. Tagged areas include Company, Ai Data, Edge AI, AI Compiler, and MLIR.

Roofline’s developer surface includes engineering blog, support, signup flow, and 7 more developer resources.

12.2/100 emerging ▬ flat Agent 0/100 human only saas Full breakdown ↓
scored 2026-09-08 · rubric v0.20.0
AccessApproval
0 APIs
CompanyAi DataEdge AIAI CompilerMLIRIREEMachine-LearningModel DeploymentEmbeddedNPUQuantizationOn-Device LLMDeveloper ToolsGermany

Kin Score

Kin Score Kin Score How this is scored →
scored 2026-09-08 · rubric v0.20.0
Create-or-Update Ergonomics could not be measured. We hold no machine-readable contract for this provider to read, so there is nothing to measure a write surface against. Excluded rather than scored zero: never-measured and measured-empty are different facts. Publishing an OpenAPI is what makes this facet — and several others — scorable at all.
Improve this rating by publishing the missing artifacts — every area above can be raised, and the full rubric is at apis.io/rating/. Every facet and dimension name above is a link: it opens that measurement's own page — what it means, the exact checks that feed it, how the whole catalog distributes on it, and the providers at the top of it. This rating is computed from github.com/api-evangelist/roofline: open an issue to ask a question, or submit a pull request to add artifacts. Submit an artifact on GitHub — free → Manage your own listing — the Influence plan, $499/mo →

Pricing Plans 1

Published pricing tiers and plan structures.

Security Posture 1

Authentication, domain security, vulnerability disclosure, and trust-center signals.

Roofline Domain Security

TLSv1.3 · HSTS

SECURITY

Resources

Get Started 1

Portal, sign-up, and the first successful call

Agent Surfaces 1

MCP servers, agent skills, and machine-readable catalogs

Build 2

SDKs, sample code, and the tooling you integrate with

Access & Security 1

Authentication, authorization, and security posture

Operate 1

Status, limits, changes, and where to get help

Commercial 2

Pricing, plans, and the legal terms of use

Company 2

The organization behind the API

Source (apis.yml)

apis.yml Raw ↑
aid: roofline
name: Roofline
description: 'Roofline (RooflineAI GmbH, Köln, Germany) is a 2024 RWTH Aachen University spin-off building a retargetable
  AI compiler stack for edge deployment. Its three products are an SDK that compiles models ahead of time from any framework
  — the product page shows a single Python call, attic.compile(model, (input,)) — a lightweight C runtime with Python bindings
  that executes the compiled model and orchestrates heterogeneous SoCs, and a Performance Dashboard fed by a roofbench command-line
  tool running nightly benchmarks across hundreds of models for latency, throughput and memory footprint. The stack is built
  on open-source MLIR and IREE with proprietary hardware-specific optimization on top, and targets NPUs, MCUs and heterogeneous
  SoCs with quantization (INT8, blocked, symmetric, asymmetric, dynamic), dynamic shapes for on-device LLMs and mixture-of-experts
  support. Published work includes NXP eIQ Neutron NPU LLM enablement, Vulkan 1.3 and Arm SVE support in IREE, and day-0 support
  for Liquid AI LFM2. Roofline is a strategic partner of the Edge AI Foundation and raised a seed round from Serena and First
  Momentum Ventures. It sells compiler licenses to hardware vendors and embedded software teams: there is no developer program,
  no public API, no documentation host and no registry-published package. Every product page ends at a Contact Sales form,
  and the only machine-readable document Roofline publishes anywhere is an llms.txt on its marketing site.'
url: https://raw.githubusercontent.com/api-evangelist/roofline/refs/heads/main/apis.yml
deliveryModel:
  model: saas
  open_source: false
  commercial: true
  callable_host: false
  label: Hosted service · you call their endpoint
  confidence: medium
  source:
  - pricing
  generated: '2026-08-28'
  method: derived
accessModel:
  pricing: unknown
  onboarding: approval
  trial: false
  try_now: false
  public: false
  label: Requires approval
  confidence: low
  source:
  - security
  generated: '2026-09-02'
  method: derived
image: https://cdn.prod.website-files.com/691c3b22c31ce2676bca90d5/691c3b22c31ce2676bca9100_roofline.svg
x-type: company
x-source: serena-portfolio
x-backed-by:
- serena
x-tier: profiled
x-tier-reason: sales-gated-compiler-sdk-no-developer-program
specificationVersion: '0.23'
created: '2026-08-17'
modified: '2026-08-17'
tags:
- Company
- Ai Data
- Edge AI
- AI Compiler
- MLIR
- IREE
- Machine-Learning
- Model Deployment
- Embedded
- NPU
- Quantization
- On-Device LLM
- Developer Tools
- Germany
tags_raw:
- Company
- Ai Data
- Edge AI
- AI Compiler
- MLIR
- IREE
- Machine Learning
- Model Deployment
- Embedded
- NPU
- Quantization
- On-Device LLM
- Developer Tools
- Germany
apis: []
maintainers:
- FN: Kin Lane
  email: kin@apievangelist.com
- FN: APIs.json
  email: info@apis.io
common:
- type: Website
  url: https://www.roofline.ai/
- type: Blog
  url: https://www.roofline.ai/news
- type: Support
  url: https://www.roofline.ai/contact-us
- type: SignUp
  url: https://www.roofline.ai/contact-us
- type: GitHubOrganization
  url: https://github.com/RooflineAI
- type: PrivacyPolicy
  url: https://www.roofline.ai/privacy-policy
- type: LLMsTxt
  url: llms/roofline-llms.txt
- type: Packages
  url: packages/roofline-packages.yml
- type: Plans
  url: plans/roofline-plans-pricing.yml
- type: DomainSecurity
  url: security/roofline-domain-security.yml
x-enrichment:
  date: '2026-08-17'
  status: minimal
  artifacts_added: 5
  pass: local-v1
  note: 'Full contract discovery ran and missed: no OpenAPI, Swagger, GraphQL, AsyncAPI, MCP endpoint or A2A agent card exists
    on any host Roofline controls, and docs/api/app/developer/mcp/sdk.roofline.ai are all NXDOMAIN. Certificate-transparency
    enumeration of %.roofline.ai surfaced 13 names, all internal engineering infrastructure; the only reachable one, cpu-demo,
    is a self-hosted Coder v2.31.9 deployment whose API belongs to coder.com and was therefore not attributed to Roofline.
    Real finds: a published llms.txt (200, 779 bytes) and a probed domain-security posture. Real absences recorded: zero registry-published
    packages despite three shipping products, and zero published plans behind a two-tier Contact Sales license model.'
x-coverage:
  state: gated
  reason: sales-gate
  detail: All three Roofline products — SDK, Runtime and Performance Dashboard — end at the same sentence, "Just reach out
    through our Contact Us form", and the only "view docs" link on those pages resolves to github.com/RooflineAI, an org whose
    own profile README disclaims its repos as "not a Roofline AI product".
  evidence:
  - url: https://www.roofline.ai/product-sdk
    status: 200
  - url: https://www.roofline.ai/contact-us
    status: 200
  - url: https://www.roofline.ai/pricing
    status: 404
  - url: https://www.roofline.ai/openapi.json
    status: 404
  - url: https://www.roofline.ai/.well-known/agent-card.json
    status: 404
  - url: https://docs.roofline.ai/
    status: 0
  checked: '2026-08-17'

Work with this as data

Every provider here is available over the APIs.io API and to AI agents over MCP.

MCP server

One button, every client — Claude, Cursor, VS Code and the rest.

https://apis.io/mcp

Tools for providers

9 MCP tools reach this
  • find_providersBrowse and filter every provider in the catalog.
  • get_provider_artifactsEvery artifact this provider publishes, grouped by type.
  • get_provider_operationsEvery operation across all of their OpenAPIs — one call instead of parsing every spec.
  • get_provider_toolsEvery MCP tool they ship, with the operation each wraps.
  • get_provider_evidenceHow each part of their score was established. Free — the basis for a claim should not sit behind it.
  • get_provider_ratingPRO — composite, band, trend and facet scores.
  • apis_io_searchSTART HERE — APIs, providers and tags for one query, each with its total.
  • resolveTurn a domain, URL or GitHub org into the provider it belongs to.
  • find_cohortsEvery scored population of providers in the catalog.
All 92 tools →

Call it yourself

curl for this page
This provider
curl "https://apis.io/api/v1/providers/roofline"
All providers
curl "https://apis.io/api/v1/providers?limit=25"
Every operation they expose
curl "https://apis.io/api/v1/providers/roofline/operations?limit=25"
How their score was established
curl "https://apis.io/api/v1/providers/roofline/evidence"

Discovery needs no key. Ratings and market analysis are Pro.

Get an API key

Free tier, no form to fill in. Signing in shares your email address with us — we store it to create your key and to recognise you if you sign in with another provider. See our Privacy Policy and Terms.

A second provider on the same verified email joins the account you already have.