Pricing built for you

Onyx pricing reflects how the system actually works: databases, operations, realtime search, and AI-driven workflows — without token math in your core plans.

Starter

Free

For learning, evaluation, early prototypes, and hobbyists.

  • 1 database
  • 1 GB storage
  • 1M reads or writes/mo
  • 1,000 AI workflow runs/mo
  • Access to all Onyx Standard models
Start building

Pro

$49 / month

For production applications and small teams.

  • Up to 5 databases
  • 10 GB storage
  • 10M reads or writes/mo
  • 6,000 AI workflow runs/mo
  • Access to all Onyx Pro models, Codex-mini and Sonnet
  • Saved workflows and reusable outputs
Upgrade to Pro

Team

$199 / month

For teams building and operating high traffic applications.

  • Unlimited databases
  • 100 GB storage
  • 50M reads or writes/mo
  • Realtime search at scale
  • 50,000 AI workflow runs/mo
  • Access to all Onyx Pro models (Codex & Sonnet)
  • Shared dashboards and workflows
Start Team

Enterprise

Custom

For organizations requiring isolation, scale, and control.

  • Dedicated Onyx hardware
  • Isolated control plane and data plane
  • Personal integration support
  • SSO / RBAC
  • Audit logs
  • Use your own AI provider keys
  • Priority support and onboarding
Talk to us

Advanced Reasoning models

Onyx includes access to specialized database tuned models that can do most data centric tasks

Advanced reasoning external models — such as Codex-Max and Anthropic Opus can be used in custom agent flows — they have a higher cost profile and are opt-in only. When enabled, they are billed separately based on actual token usage.

Explicit workflow-level opt-inFull usage visibility

Pricing FAQ

What is an AI workflow run?
A workflow run represents a complete AI-assisted cycle: intent interpretation, query generation, relationship resolution, execution, and output creation.

Are Codex and Sonnet considered premium?
No. Codex and Anthropic Sonnet are included as part of Onyx Pro models. Only Codex-Max and Anthropic Opus are considered premium external usage.

How are premium models billed?
Premium models are billed separately based on actual token usage when explicitly enabled.

Bring your own API key?
Supported on Enterprise plans.

Start free. Scale deliberately.

No forced upgrades. No surprise behavior.

Start building