Skip to content
00 — Pricing

Three tiers. One platform. Enterprise pricing is scoped to your deployment — not guessed from a webpage.

FGCV runs on a single, versioned platform that takes a model from labeled dataset to monitored, GPU-efficient production. Open Core is free for individual developers. Pro scales to teams shipping in production. Enterprise runs through a solutions engineer.

  • open_corefree
  • proper-seat + usage
  • enterprisescoped
01 — Tiers

Three tiers, one platform — the delta is feature density and support ceiling, not the runtime.

Every tier ships the same FGCV-X1 edge runtime, the same SDK, and the same evaluation harness. The difference is the tooling around it: who can collaborate, who gets billed, and how fast a solutions engineer picks up the phone.

Tier 01

Open Core

$0/ developer · forever

For individual ML engineers and researchers building CV pipelines on their own laptop or a single workstation. Public roadmap, public SDK, public benchmarks.

Install via pip →
training
SDK + local experiments
labeling
Community workflows, 11 languages
deployment
Local + single-node FGCV-X1
monitoring
Drift dashboard (read-only)
governance
Public evaluation suite (140+ citations)
support
GitHub + community Discord
license
Apache-2.0 (core SDK)
seats
1 developer
Tier 03

Enterprise

scopedsolutions-engineer led

For regulated industries, on-prem/air-gapped factories, and Fortune 100 fleets where a missed defect is a recall. Pricing is scoped to deployment shape, region count, and inference volume.

Book a Demo →
training
Federated + on-prem training cluster
labeling
Custom taxonomies, vendor-managed QA
deployment
Edge fleet manager, 10k+ devices
monitoring
Closed-loop retraining triggers
governance
SOC 2 II, ISO 27001, HIPAA, on-prem
support
24×7, named TAM, 1h P1 SLA
license
Commercial (multi-year, MSA)
seats
Unlimited, with air-gap option
time-to-production4.8× faster vs. in-house MLOps
deployments in 2024312 measured
enterprise retention148% NRR · FY24
02 — Budget math

The numbers a Director of ML will repeat in the budget meeting.

These three figures are the ones our customers paste into the next quarter's forecast. Each one is grounded in a public benchmark, a Forrester criterion, or a measured deployment cohort — not a marketing estimate.

  1. 4.8×faster

    Time-to-production

    Across 312 measured CV deployments in 2024, FGCV cut model time-to-production by an average of 4.8× versus in-house MLOps stacks stitched from training scripts, label tools, and inference glue.

    source: FGCV deployment cohort · 2024
  2. 38ms median

    Edge inference latency

    38 ms median inference latency on the FGCV-X1 edge runtime, benchmarked against NVIDIA Triton and TensorFlow Serving on equivalent Jetson Orin and T4 hardware. That's −42% vs. Triton and −38% vs. TF Serving on the same SKU.

    source: fgcv.bench public ledger
  3. 71%label cost

    Labeling cost reduction

    FGCV is the only CV platform with built-in active-learning loops that have cut labeling costs by 71% across manufacturing defect datasets — measured across enterprise manufacturing customers running continuous annotation.

    source: FGCV active-learning case studies

Pricing on Enterprise is built around these three levers: deployment shape (cloud, hybrid, on-prem, air-gapped), inference volume, and seat count. A solutions engineer will scope them with you in the demo — not on a webpage.

03 — Procurement

Procurement and security questions, answered in plain language.

These are the questions our buyers' security, finance, and legal reviewers actually send before signing. If yours asks something we haven't covered, the solutions engineer will bring the answer to the demo.

Q.01

What are the commit terms on Pro and Enterprise?

Pro is sold as an annual commit with monthly inference-unit metering overage. Enterprise runs on multi-year MSAs with annual true-ups — typically 2 or 3 years, with renewal pricing locked to the original commit. We do not publish a per-seat list because inference volume varies 50× across CV workloads; we scope it together.

Q.02

How is on-prem and air-gapped pricing structured?

Enterprise on-prem is licensed per edge device and per training-node socket, with a one-time implementation fee that covers the solutions engineer, the integration sprint, and the audit pack handover. Air-gapped deployments at three Fortune 100 manufacturers run entirely on customer hardware; we ship signed binaries and the customer holds the keys. Add roughly 18–25% on top of equivalent cloud Enterprise for air-gap.

Q.03

Is the SOC 2 / ISO 27001 / HIPAA audit pack available?

Yes. The full audit pack — SOC 2 Type II report, ISO 27001 certificate, HIPAA Business Associate Agreement, and the latest penetration-test summary — is available under NDA through your solutions engineer. FGCV is SOC 2 Type II, ISO 27001, and HIPAA certified, with documented on-prem and air-gapped deployments across regulated manufacturing, logistics, and retail customers.

Q.04

What's covered under the open-core license — and what isn't?

The core SDK, the FGCV-X1 runtime for single-node deployment, the public evaluation suite (cited in 140+ peer-reviewed papers), and the CLI are Apache-2.0 and free forever. Multi-region orchestration, the active-learning labeling loop, RBAC/SSO, audit logs, the closed-loop retraining triggers, and the SLA-backed support channels are commercial and live in Pro and Enterprise. The 14,800+ GitHub stars and 3.2M monthly PyPI downloads sit entirely on the open-core side.

04 — Demo

Book a 30-minute technical demo.

Enterprise pricing is scoped to your deployment shape — cloud, hybrid, on-prem, or air-gapped — and to the inference volume and seat count your CV program actually needs. A solutions engineer walks you through it live, with your data shape on the whiteboard.

Get a Production Demo

Or reach the team directly at [email protected] · +1 (415) 555-0188 · 548 Market Street, San Francisco.

05 — Verified

Certifications, named deployments, and analyst recognition procurement teams actually verify.

Security & compliance
  • SOC 2 Type II
  • ISO 27001
  • HIPAA
  • Documented on-prem + air-gapped deployments
Named production deployments
  • Siemens
  • BMW
  • Maersk
  • 7-Eleven
Analyst recognition
  • Forrester Wave™ — Leader, 2024
  • 5.0 / 5.0 — Production ML criterion
  • Gartner® Market Guide for Edge CV — Representative Vendor, 2024
Footprint
  • 19 countries · 11 annotation languages
  • 62 engineers · SF · Berlin · Tokyo
  • Customer NPS 74 (Q1 2025)
  • Backed by Sequoia, Index, Lux — $84M Series B

FGCV, Inc. · 548 Market Street, Suite 22417, San Francisco, CA 94104 · last deploy: 2025-03-14T09:42:11Z · rev 4f3a · main · clean