Simple, predictable pricing

Scale agent quality from first trace to enterprise review.

Start free, then pay for the telemetry and evaluation volume your team actually uses. Plans grow from single-developer observability to secure organization-wide AI quality programs.

Hobby

Free

forever

Perfect for individual developers getting started with LLM and OTLP observability.

Usage

No metered overage

  • 50,000 monthly events included
  • 30 days data access
  • Up to 2 organization users
  • Up to 2 projects
  • Community support
Start free

Core

$29

/ month

Great for small teams with unlimited users and more data access.

Usage

+ $8-6 / 100k events, graduated pricing

  • 90 days data access
  • Unlimited users
  • Unlimited projects
  • First 100,000 events included
  • Usage-based alerts
  • Email and chat support
Start trial

Pro

$199

/ month

Popular

For scaling projects that need advanced evaluations, compliance, and longer access.

Usage

+ $8-6 / 100k events, graduated pricing

  • Everything in Core
  • 3 years data access
  • Advanced evaluations
  • Data retention management
  • SOC2 and ISO27001 compliance
  • High rate limits
Start trial

Team

$499

/ month

Organizational and security controls for larger teams.

Usage

+ $8-6 / 100k events, graduated pricing

  • Everything in Pro
  • Enterprise SSO
  • SSO enforcement policy
  • Fine-grained RBAC roles
  • Audit logs support
  • Dedicated Slack support
Start trial

Enterprise

$2,499

/ month

For large enterprises requiring custom SLAs, dedicated support, and admin APIs.

Usage

Custom volume discounts available

  • Everything in Team
  • Admin API access
  • SCIM provisioning API
  • Uptime and support SLAs
  • Dedicated support engineer
  • Custom volume pricing
Contact sales

Metered cloud usage

Built for teams whose agents keep getting more traffic.

Provon pricing separates platform capabilities from usage. Pick the control plane your team needs, then let event volume grow without forcing every teammate into seat-based math.

Events included

100K events

Core and above include the first 100K events every month.

Metered overage

$8-6 / 100K

Graduated cloud event pricing keeps scale predictable.

Retention controls

30 days - 3 years+

Move from quick debugging to compliance-grade history.

Compare plans

Choose your control plane.

Need custom volume pricing?
CapabilityHobbyCoreProTeamEnterprise
Workspace
Organization users2UnlimitedUnlimitedUnlimitedUnlimited
Projects2UnlimitedUnlimitedUnlimitedUnlimited
Included monthly events50K100K100K100K100K+
Evaluation
Traces, logs, and metricsIncludedIncludedIncludedIncludedIncluded
Findings and online evalsBasicIncludedAdvancedAdvancedAdvanced
Datasets and personasBasicIncludedIncludedIncludedIncluded
Data and security
Data access window30 days90 days3 years3 yearsCustom
RBAC and audit support--Compliance controlsFine-grained RBACAdmin APIs
SSO---SAML / OIDCSAML / OIDC + SCIM
Support
Support channelCommunityEmail + chatPriorityDedicated SlackDedicated engineer
SLA----Custom SLA

Startups

Founder-friendly

Tell us about your stage and usage profile.

Open source

Credits available

Qualified OSS teams can ask for cloud credits.

Education

Flexible

Student and research teams can request special access.

FAQ

Pricing questions.

How is event-based pricing calculated?

Provon meters telemetry and evaluation activity as events. Paid plans include a monthly event allowance, then apply graduated pricing for additional cloud events.

Can we start free and upgrade later?

Yes. Hobby is built for individual projects. Teams can upgrade as they need more users, projects, retention, security controls, and support.

Do you support self-hosted or enterprise deployments?

Enterprise customers can discuss custom deployment, support, procurement, volume discounts, and admin API requirements with the Provon team.

What is included in retention management?

Paid tiers unlock longer data access windows and retention controls for traces, logs, metrics, findings, datasets, and evaluation history.

Ready to instrument agent quality?

Start with real traces, then turn them into evals, findings, and regression datasets.