JFrog pricing

About

JFrog is the company behind Artifactory, the binary/artifact repository that underpins software delivery pipelines from individual developers up through Fortune 100 enterprises — its pricing page names Airbus, American Express, Box, Fidelity, and Mercedes among customers, and claims usage at “over 80% of the Fortune 100.” The platform has expanded outward from binary/artifact management into a broader “software supply chain” suite: Xray for security scanning (SCA, secrets detection, license/vulnerability scanning), Distribution for release management, a container registry, and — most recently — AI/ML model lifecycle tooling (Model Registry, AI Catalog, Shadow AI Detection, Feature Store) that manages trained models the same way Artifactory manages binaries.

JFrog (Nasdaq: FROG) sells the same tier structure through two parallel deployment paths — a fully managed SaaS product billed monthly against a storage-and-transfer consumption meter, and a Self-Managed (self-hosted) product billed annually against a server-count license — so a customer chooses the operational model without changing which product tier (Pro/Pro X, Enterprise X, Enterprise +) they buy. JFrog positions itself directly against Sonatype (artifact management + security), GitLab (integrated source control/CI), and Snyk (point-solution security), arguing on its own “Compare JFrog” page that none of the three combines binary-level insight with full-lifecycle security and AI governance in one platform.

Pricing summary : How JFrog prices SaaS vs. Self-Managed software supply chain tiers

JFrog uses a hybrid, dual-deployment model with 4 dimensions:

1. Deployment path: SaaS (monthly, metered against a storage+transfer consumption pool) or Self-Managed (annual, licensed per server count) — the same three tier names (Pro/Pro X, Enterprise X, Enterprise +) exist on both paths at different prices and units.
2. Base capacity per tier: SaaS Pro includes 25 GB base consumption (storage+transfer combined), Enterprise X includes 125 GB, Enterprise + is custom — each with “additional consumption available” beyond the base. Self-Managed ships 1 server (Pro X), 3 servers (Enterprise X), or 6 servers (Enterprise +) instead of a consumption meter.
3. List price vs. active price: SaaS Pro’s list price is $150/month, but a “Limited Time Offer” cuts it to $50/month (Save $100/mo) — the discounted price is disclosed as available only through June 4, 2027 on the checkout page, not the main pricing page.
4. Paid add-on bundles: Security Bundles (Unified/Ultimate) and MLOps Bundles (Unified/Ultimate) layer optional per-feature charges — e.g., JFrog AI Catalog, Runtime Impact, Transitive Contextual Analysis — on top of either deployment path’s base tier. Each bundle is gated by a developer-count base (200 devs for Security; 200/500 devs for MLOps Unified/Ultimate), and most add-on rows show a bare ”$” rather than a disclosed rate.

What makes this different: JFrog doesn’t pick one billing axis — it prices the identical three-tier product ladder twice (once as SaaS consumption, once as an annual self-managed server license), then stacks a second, mostly-undisclosed consumption dimension (security/MLOps bundles) on top of whichever path a customer chooses.

Pricing by product

JFrog Platform (SaaS plans)

JFrog Platform (Self-Managed plans)

Sales motions across products: PLG / self-serve for SaaS Pro (checkout at jfrog.com/buy-now/pro/, or AWS/GCP/Azure Marketplace); sales-led (“Request A Quote” / “Contact Us”) for every other tier on both SaaS and Self-Managed, including Self-Managed Pro X.

JFrog Bundles — Security & MLOps add-ons layered on top of a base tier

Both deployment paths share the same optional add-on structure, shown on the pricing page independent of the SaaS/Self-Managed toggle:

None of the ”$” (paid add-on) cells disclose a specific rate anywhere in the captured surfaces — pricing for those line items requires contacting sales.

Hidden costs : SaaS-vs-Self-Managed TCO and the undisclosed bundle add-ons

JFrog’s advertised headline prices — $50/month for SaaS Pro, $27,000/year for Self-Managed Pro X — describe only the base tier. The two real-world levers that move a JFrog bill the most are (1) which deployment path you pick, since SaaS and Self-Managed price the same tier ladder on completely different economics, and (2) whether you attach a Security or MLOps bundle, whose per-feature rates are undisclosed anywhere in the public funnel.

Archetype 1 — SaaS vs. Self-Managed, 3-year cost of an Enterprise X deployment

A mid-market engineering org comparing the two deployment paths for the same Enterprise X tier sees a very different sticker price depending on who hosts it:

Self-Managed’s 3-year list price runs roughly 4.5× higher than the equivalent SaaS tier’s list price alone — before SaaS’s own per-GB consumption overage is added on top — because Self-Managed customers are paying for the server license and absorbing their own hosting, patching, and uptime engineering, work JFrog’s SaaS fee otherwise covers. A team that can’t yet size its consumption footprint should model the SaaS overage curve (JFrog’s published Enterprise X overage rate is $1.45/GB storage and $0.45/GB transfer beyond the base pool, the same per-GB storage pricing mechanic used across the corpus) before assuming SaaS is automatically cheaper at scale — this is exactly the kind of AI cost unpredictability and bill-shock risk that a combined consumption meter creates for buyers.

Archetype 2 — Stacking a Security Bundle and an MLOps Bundle on top of Enterprise X

A 300-developer organization that wants Enterprise X’s DevSecOps baseline plus AI-model governance has to add two separate bundles, both gated by headcount rather than usage, and both showing undisclosed rates:

Because neither bundle’s per-feature rate is published, the same nominal $950/month base tier can land anywhere from roughly $1,000/month to several times that once Security and MLOps add-ons are priced in — the “Bundle & Save” framing on the pricing page markets the packaging, not the number a buyer needs to build a real budget.

Want to estimate your own JFrog bill? Use the JFrog pricing calculator to model your monthly cost based on storage, transfer, and deployment path.

Pricing evolution : From flat GB caps to a combined consumption pool plus a metered AI add-on

Cadence

Tracked range: 2022-01–2026-09. The 2022 Q1–Q3 and 2023 Q1 windows were directly Wayback-verified stable (0 price changes, 0 SKU additions) beyond what’s listed above. The 2023 Q2–2024 Q1 window falls in a gap between direct archive capture and analyst reporting and is not independently verified — treated as unknown rather than assumed stable, including the exact date the SaaS consumption model was repackaged from separate storage/transfer/CI-CD-minute meters into today’s single combined pool.

Notable changes

• 2022-11 — SaaS “Enterprise” renamed “Enterprise X,” price rising from $699/month to $1,199/month (+72%); Free tier shrinks from unlimited users/10 GB transfer to 5 users/5 GB transfer (Wayback captures of jfrog.com/pricing).
• 2024-06-25 — JFrog completes its $230 million acquisition of Qwak AI, the MLOps platform that becomes the technical foundation of JFrog ML (JFrog investor press release).
• 2025-01-01 — Self-Managed Pro X list price rises about 15%; Self-Managed Enterprise X rises about 6% (Barclays equity research, reported via Nasdaq/TipRanks, January 2, 2025).
• 2025-03-04 — JFrog launches JFrog ML, its first Qwak-derived product, positioning the company as an “AI System of Record” — Model Registry, AI Catalog, and AI-governance tooling ship bundled inside existing SaaS/Self-Managed tiers rather than as a standalone SKU (JFrog/BusinessWire press release).
• 2025-04-02 — Barclays reports JFrog raised its SaaS Enterprise X list price by roughly 27% and added a metered ML credits add-on ($1.20/credit) to both SaaS and Self-Hosted plans — the first explicit usage-based charge tied to AI/ML workloads on the platform.
• 2025-11-13 — JFrog unveils Shadow AI Detection at swampUP Europe, an AI-governance feature that discovers and catalogs unauthorized (“shadow”) AI model and API usage, shipped inside JFrog AI Catalog.
• 2026-01-01 — Self-Managed Enterprise X rises about 6% (vs ~5% in 2025); Self-Managed Pro X holds flat (vs a 15% hike in 2025) — continuing an annual January-1 self-hosted repricing cadence (Barclays equity research).

The SaaS Enterprise X price hike and ML-credit add-on in detail

The April 2025 change is the clearest evidence in JFrog’s history that AI/ML workloads are becoming their own billing dimension rather than riding for free inside the existing storage-and-transfer meter. Barclays’ research desk — not JFrog’s own marketing — is the source for both halves of this change: a roughly 27% list-price increase on the SaaS Enterprise X tier, announced with no corresponding blog post or changelog entry, and a new $1.20-per-credit charge for “ML credits” layered on top of both the SaaS and Self-Hosted plans. Read together with JFrog ML’s March 2025 launch (Model Registry, AI Catalog, and later Shadow AI Detection all shipping as bundled features rather than new SKUs), the sequence suggests JFrog’s strategy is to keep the feature bundled into the base tier for adoption, while introducing a usage meter underneath it once the feature has customers to bill — the same rate-limit-to-credits billing shift other AI-adjacent products have gone through as usage patterns firm up. Because Barclays flagged the possibility that the changes apply to new customers only, existing customers may not see the increase reflected in their next renewal — a distinction JFrog’s own pricing page does not disclose.

What’s unique : Dual-deployment pricing and AI tooling bundled into the artifact ladder

1. The same three-tier ladder is priced twice, on two different economics. Most vendors either sell SaaS or self-hosted; JFrog sells the identical Pro/Enterprise X/Enterprise+ naming on both, but SaaS meters monthly consumption while Self-Managed licenses an annual server count. A buyer chooses the operational model — who patches, who scales, who owns uptime — independently of which feature tier they need, a packaging choice distinct from most other developer tools pricing in the corpus.

2. AI/ML lifecycle tooling ships bundled into the existing artifact ladder, not as a new product line. JFrog ML — built from the 2024 Qwak AI acquisition — puts a “Comprehensive ML Model Registry” inside the base Pro tier already, while more advanced MLOps capabilities (Experiment Tracking, Model Monitoring, Feature Store) sit behind the same MLOps Bundle add-on structure used for security features. Competing MLOps platforms typically sell model-lifecycle tooling as its own SKU; JFrog treats it as a feature of the artifact repository it already sells.

3. Add-on bundles are gated by developer headcount, not by consumption. The Security and MLOps Bundles don’t meter GB or API calls — they’re priced against a “Base Bundle: 200 Devs” (or 500 for MLOps Ultimate) headcount, an unusual hybrid where an add-on’s entry price scales with team size rather than usage volume, layered on top of a base tier that itself is metered by consumption or server count.

4. The one confirmed usage-metered AI SKU is documented by analysts, not by JFrog’s own pricing page. Per Barclays equity research, JFrog added a $1.20-per-credit charge for ML workloads in April 2025 — a real, dated, usage-based line item for AI/ML consumption that doesn’t appear anywhere in JFrog’s own public pricing copy, underscoring how much of JFrog’s actual AI monetization sits behind the “Contact Us” wall.

5. A conversational AI sales assistant is embedded directly in the pricing page itself. “Ask Arty” sits inline on jfrog.com/pricing as a chat widget for real-time pricing questions — turning the pricing page into an active sales-qualification surface rather than a static price list, a mechanic distinct from JFrog’s separate Product Selector qualifier tool.

Strengths & weaknesses

Billing UX : Toggles, checkout flow, and qualifier tools on jfrog.com/pricing

• SaaS / Self-Managed toggle — a two-tab switch at the top of jfrog.com/pricing/ that swaps the entire tier grid (names, prices, and included features) between monthly SaaS consumption pricing and annual Self-Managed server-count pricing, without changing the page URL.
• Buy Now cloud-provider picker — the jfrog.com/buy-now/pro/ checkout presents a radio choice between AWS, Google Cloud, and Microsoft Azure, a 13-region server-location dropdown, and a hostname field (<name>.jfrog.io) before the CONFIRM button activates; it also surfaces the AWS Marketplace path as an alternative with its own pricing.
• Feature Comparison Spotlight — collapsible category rows (Platform Management & Consumption, DevOps & Binary Management, Application & Platform Security, AI Gov & Automation, IoT & Edge Device Management) that expand a shared multi-column tier comparison table, mirrored under both the SaaS and Self-Managed toggle states.
• JFrog Product Selector — a standalone 5-question qualifier at jfrog.com/product-selector/ (developer-count slider from 0 to 10,000, a security-concerns checklist, an IAM-requirement yes/no, a multi-datacenter-needs multiple-choice, and an implementation-bandwidth yes/no) that routes prospects to a recommended package instead of showing price directly.
• “Ask Arty” AI sales assistant — a chat widget embedded on the pricing page itself for real-time pricing questions (“Arty, our AI sales rep, requires cookies to work”).

Strategic wins : Why JFrog’s dual-path, bundled-AI packaging worked

1. Pricing the same tier ladder twice instead of forcing a deployment decision

By selling identical Pro/Enterprise X/Enterprise+ naming on both SaaS and Self-Managed, JFrog lets security- or compliance-constrained buyers choose self-hosting without losing feature parity, and lets everyone else default to SaaS for lower operational overhead. This mirrors the packaging lesson in our guide to usage-based pricing implementation — separate the mechanic (what’s included) from the delivery model (who runs it) so one doesn’t force a re-negotiation of the other.

2. Folding AI/ML lifecycle tooling into the existing meter instead of a parallel bill

Rather than spinning JFrog ML into its own priced product after the Qwak acquisition, JFrog shipped Model Registry and AI Catalog as line items inside the tiers customers already buy. That avoids the adoption friction of a brand-new SKU during the sensitive early period when a company is still proving out AI/ML governance value — a sequencing our FinOps for AI cost management coverage argues is the right approach for nascent AI cost centers: bundle for adoption, meter later once usage is proven.

3. A public, if noisy, price list across all deployment/tier combinations

Even with undisclosed bundle add-ons, JFrog still publishes a full base price grid across six combinations (3 tiers × 2 deployment paths) — a level of pricing transparency that many competing DevSecOps point-solutions (Snyk, Sonatype) don’t match at the enterprise tier, giving buyers a real anchor before they ever talk to sales.

Areas to improve : Where JFrog’s packaging still hides the real number

1. Undisclosed bundle rates create a shadow contact-sales wall inside a self-serve page

Every Security and MLOps Bundle line item that shows a bare ”$” — JFrog AI Catalog on Unified Security, Runtime Impact on both tiers, the entire MLOps Bundle rate card — forces a sales conversation that the rest of the page is designed to avoid. Fix: publish at least an indicative starting rate or range for each bundle, the way JFrog already does for the base tiers; even a “starting at” figure would let self-serve buyers self-qualify before engaging sales.

2. Consumption metering penalizes the platform’s own core use case

Because CI/CD pipelines pulling artifacts count as “transfer” the same way a human download does, teams that lean hardest into JFrog’s flagship CI/CD integration are the ones most likely to blow through their base consumption pool — the exact pattern documented in third-party TCO teardowns of the platform. Fix: exclude internal, same-account CI/CD artifact pulls from the billable transfer meter, or offer a dedicated CI/CD allowance the way the 2022-era pricing page once did with its separate CI/CD-minutes line.

3. No visible price-change notice mechanism for existing customers

The annual January 1 Self-Managed increases (documented only by Barclays equity research, not by JFrog’s own blog or changelog) leave existing customers with no public record of what changed or when. Fix: publish a dated pricing changelog — mirroring the usage-invoicing and billing-cycle guidance on proactive change communication — so renewing customers aren’t the only ones who ever see the new number.

Key takeaways

1. A single tier ladder can be priced twice without confusing buyers, as long as the naming stays identical. JFrog reuses Pro/Enterprise X/Enterprise+ across both SaaS and Self-Managed rather than inventing separate brand names per path, so a customer switching deployment models keeps a stable mental map of what they’re buying.
2. Bundling a new product line into existing tiers avoids the “new SKU” adoption tax. JFrog ML shipped as line items inside tiers customers already own rather than as a fresh contract negotiation, letting AI/ML governance features get initial usage before JFrog introduces separate metering.
3. Consumption meters that combine two cost drivers into one number sacrifice forecastability for simplicity. JFrog’s single storage+transfer pool is easy to state on a pricing page but hard for a buyer to model, because storage and transfer usually grow on different curves for the same team.
4. Analyst research can surface pricing changes a vendor never announces. Every dated price increase in JFrog’s 2025-2026 history in this analysis comes from Barclays equity research, not from a JFrog blog post or changelog — a reminder that “no announcement” doesn’t mean “no change.”
5. Gating add-on bundles by headcount instead of usage creates a hybrid pricing signal that’s easy to overlook. JFrog’s Security and MLOps Bundles scale their entry price with developer count rather than consumption volume, a mechanic worth naming explicitly rather than lumping into “enterprise pricing.”

UBP implications

1. AI/ML features are increasingly launched bundled, then metered later once adoption is proven. JFrog ML’s March 2025 launch inside existing tiers, followed a month later by a documented usage-metered ML-credit charge, is a template other platforms bolting AI onto an existing product are likely to repeat: ship the feature free-riding on the base subscription, then introduce the meter once usage data justifies it.
2. Deployment-model choice is becoming its own pricing axis, independent of feature tier. JFrog’s SaaS-vs-Self-Managed split — same tier names, radically different unit economics (monthly consumption vs. annual server license) — is a pattern usage-based pricing teams designing for both cloud-native and compliance-constrained buyers should study directly.
3. Undisclosed add-on rates inside an otherwise-public price list are a rising pattern worth tracking across the corpus. JFrog’s bare ”$” symbols for Security/MLOps bundle features sit inside a page that is otherwise unusually transparent about base-tier pricing — a hybrid transparency posture that may become more common as vendors try to keep headline prices legible while still gating the newest (often AI-related) capabilities behind sales.

Sources

• JFrog pricing page (accessed 2026-09-01)
• JFrog Buy Now — Pro checkout (accessed 2026-09-01)
• JFrog Start Free / trial page (accessed 2026-09-01)
• JFrog Product Selector (accessed 2026-09-01)
• Compare JFrog (accessed 2026-09-01)
• JFrog documentation (accessed 2026-09-01)
• JFrog release notes (accessed 2026-09-01)
• JFrog blog (accessed 2026-09-01)

Bottom line

JFrog turns a single artifact-management tier ladder into two separate pricing conversations — SaaS consumption and Self-Managed server licensing — then layers a mostly undisclosed set of Security and MLOps bundles, and now a metered AI-credit charge, on top of whichever path a buyer picks; the result is a page that looks unusually transparent at the base tier and unusually opaque one click deeper.

Want to compare JFrog against other developer-tools and DevSecOps pricing? Browse the pricing blueprint.

Pricing timeline : Major events on a vertical axis

Each milestone below corresponds to a public pricing change, product launch, or material adjustment. Major events use a filled marker; minor adjustments use a faded one.

Current ladder: SaaS Pro discounted to $50/mo; Self-Managed Pro X flat at $27,000/yr

Sep 2026

As captured: SaaS Pro lists at $150/month with a 'Limited Time Offer' discounting it to $50/month through June 4, 2027; Enterprise X starts at $950/month; Enterprise+ is custom. Self-Managed runs $27,000/year (Pro X), $51,000/year (Enterprise X), and custom (Enterprise+). The exact date the SaaS ladder was repackaged from separate storage/transfer/CI-CD-minute meters (2023) to today's single combined consumption pool is not preserved in accessible archives and is marked unknown rather than guessed.

Self-Managed Enterprise X rises again; Pro X holds flat

Jan 2026

Barclays equity research reports JFrog raised its Self-Managed Enterprise X price about 6% effective January 1, 2026 (versus about 5% in 2025), while the Self-Managed Pro X price held flat (versus a 15% hike in 2025) — continuing an annual January-1 self-hosted repricing pattern. No dollar figures were disclosed.

Shadow AI Detection unveiled inside JFrog AI Catalog

Nov 2025

JFrog introduces Shadow AI Detection at its swampUP Europe conference — an AI-governance feature that discovers and catalogs unauthorized ('shadow') AI model and API usage — shipped as part of JFrog AI Catalog rather than as a newly priced SKU. Source: JFrog press release / DevOps Digest coverage, November 13, 2025.

SaaS Enterprise X price raised ~27%; metered ML-credit add-on introduced

Apr 2025

Barclays equity research (reported via financial press) found JFrog's pricing page showed the SaaS Enterprise X tier raised roughly 27%, and a new option to purchase additional 'ML credits' at $1.20/credit added to both SaaS and Self-Hosted plans — the first explicit usage-metered charge tied to AI/ML workloads on the platform. Barclays speculated the changes may apply to new customers only, given limited revenue impact from the 2025 self-hosted increases.

JFrog ML launches, bundling AI/ML lifecycle tooling into existing tiers

Mar 2025

JFrog launches JFrog ML, positioning the company as an 'AI System of Record' — Model Registry, AI Catalog, and (later) Shadow AI Detection ship as line items inside the existing Pro/Enterprise X/Enterprise+ ladder on both SaaS and Self-Managed, rather than as a standalone product SKU. Source: JFrog/BusinessWire press release, March 4, 2025.

Self-Managed Pro X and Enterprise X list prices rise

Jan 2025

Per Barclays equity research reported by financial press (Nasdaq/TipRanks, Jan 2, 2025), JFrog raised its Self-Managed Pro X list price by about 15% and Self-Managed Enterprise X by about 6%, effective January 1, 2025; analysts estimated a conservative $10M FY25 self-hosted revenue impact. No dollar figures were disclosed on the pricing page itself.

JFrog completes $230M acquisition of Qwak AI

Jun 2024

JFrog acquires Qwak AI Ltd., an MLOps platform, for $230 million — the technical foundation for what becomes JFrog ML. No pricing SKU changes accompanied the close; the product itself launched separately in March 2025. Source: JFrog investor press release, June 2024.

Enterprise renamed Enterprise X with a 72% price jump; Free tier shrinks

Nov 2022

Between the October and November 2022 Wayback captures, JFrog renamed its SaaS 'Enterprise' plan to 'Enterprise X' and raised its price from $699/month to $1,199/month (+72%); the same snapshot shows the Free tier's allowance cut from unlimited users/10 GB transfer to 5 users/5 GB transfer. Source: Wayback Machine captures of jfrog.com/pricing, Oct-Nov 2022.

Baseline SaaS ladder: Free, Pro Team, Enterprise, Enterprise+

Jan 2022

The earliest captured JFrog SaaS pricing page shows a $0/month Free tier (unlimited users, 10 GB transfer, 2 GB storage), Pro Team at $98/month, Enterprise at $699/month, and custom-quoted Enterprise+ — each metering storage and transfer as separate GB allowances, plus a CI/CD-minutes allowance. Source: Wayback Machine capture of jfrog.com/pricing, January 2022.