Community
$0
- Local runtime & open specialist models
- SDK, schemas, example policies
- Self-hosted, air-gapped friendly
Pricing — proposed
Transparent ranges, not final prices
Nothing is for sale on this site. There is no signup, checkout, or billing. The tiers below are proposed directions with documented assumptions, published so the model can be reviewed openly.
No billing implemented. Inference cost has not been measured, so usage prices are indicative and subject to change. Community tooling is intended to remain free.
Developer
Usage + monthly credits
- Hosted API for prototyping
- Initial monthly credits included
- Priced per event / agent / request
Pro
$49–$199 / month
- For developers & small teams
- Higher limits than Developer
- Range depends on usage volume
Business
$500–$5k+ / month
- Teams & growing companies
- Scales with protected assets
- Priority support (planned)
Enterprise
Annual contract
- Priced on endpoints, agents, telemetry
- Tenant isolation & audit
- Deployment support
Infrastructure · Government · OEM
Custom
- Private & air-gapped deployment
- Volume licensing for OEM
- Contractual pricing
Transparency
Assumptions behind these ranges
Pricing dimensions we are modeling
- Events analyzed — the primary metered unit for API usage.
- Protected agents — number of AI agents routed through the Guardian.
- Protected assets / endpoints — for Business and Enterprise scaling.
- Compute time — for heavier correlation and reasoning workloads.
Why Pro sits at $49–$199
Common developer-tooling price points for small teams, bounded so the range stays honest until real usage and inference costs are measured. The lower bound targets a single developer; the upper bound a small team with steady volume.
What is not included in these numbers
- Measured inference cost — pending. Until measured, usage prices are indicative only.
- Third-party infrastructure pass-through costs.
- Any implied discounts, promotions, or contractual terms.
How this evolves. Once the runtime and models are benchmarked, we will publish unit economics (cost per event, per agent) and revise these ranges accordingly. The methodology lives with the project's business and benchmark records.