Start Free. Pay Only For What You Run
Keep your stack. No long contracts—scale usage up or down.
Start Free. Pay Only For What You Run
Keep your stack. No long contracts—scale usage up or down.
Start Free. Pay Only For What You Run
Keep your stack. No long contracts—scale usage up or down.
Choose your set up
Cloud
Choose if you want us to manage the MPC network.
Self-hosted
Choose if you want to run the Stoffel Stack yourself with the same developer experience.
GENERAL MPC COMPUTE
Cloud
Use this when you’re running any Stoffel Lang program you write—analytics, ML steps, matching, key ops—without committing to a specific pre-packaged solution.
MPC Compute Units (MCUs) = seconds of compute × number of parties. Example: a 3-party run that lasts 2 seconds = 6 MCUs.
Developer
2 million MCUs
per month
Up to 5 concurrent sessions
Basic audit logs
Low per-MCU overages
Team
20 million MCUs
per month
Up to 25 concurrent sessions
SSO
Policy packs
Full audit logs
Priority support
Choose your set up
Cloud
Choose if you want us to manage the MPC network.
Self-hosted
Choose if you want to run the Stoffel Stack yourself with the same developer experience.
GENERAL MPC COMPUTE
Cloud
Use this when you’re running any Stoffel Lang program you write—analytics, ML steps, matching, key ops—without committing to a specific pre-packaged solution.
MPC Compute Units (MCUs) = seconds of compute × number of parties. Example: a 3-party run that lasts 2 seconds = 6 MCUs.
Developer
2 million MCUs
per month
Up to 5 concurrent sessions
Basic audit logs
Low per-MCU overages
Team
20 million MCUs
per month
Up to 25 concurrent sessions
SSO
Policy packs
Full audit logs
Priority support
Choose your set up
Cloud
Choose if you want us to manage the MPC network.
Self-hosted
Choose if you want to run the Stoffel Stack yourself with the same developer experience.
GENERAL MPC COMPUTE
Cloud
Use this when you’re running any Stoffel Lang program you write—analytics, ML steps, matching, key ops—without committing to a specific pre-packaged solution.
MPC Compute Units (MCUs) = seconds of compute × number of parties. Example: a 3-party run that lasts 2 seconds = 6 MCUs.
Developer
2 million MCUs
per month
Up to 5 concurrent sessions
Basic audit logs
Low per-MCU overages
Team
20 million MCUs
per month
Up to 25 concurrent sessions
SSO
Policy packs
Full audit logs
Priority support
Solutions Pricing
Use this when you want a well-defined use case that ships faster: each solution includes Lang modules, VM presets, and an app-specific SDK. Every solution lets you pick Per call or Per cluster so billing matches how you operate.
Federated Learning Aggregator
Replace the central FL aggregator with shared computation. Train one model; only the combined update leaves.
FL Aggregator
Private Queries
Analytics
Key Generation
Per call
By Round (recommended)
Pay per aggregation round.
Scale up to [N parties] and gradient size up to [X size]. Or choose MCU pricing for maximum flexibility
Per cluster
Predictable monthly usage
FL Cluster tiers include a monthly allowance of rounds (e.g., 2k / 10k / 50k) with reserved training-time concurrency.
All tiers include:
Starter (2k rounds/mo)
Flower-friendly adapter SDK
Round monitoring
Solutions Pricing
Use this when you want a well-defined use case that ships faster: each solution includes Lang modules, VM presets, and an app-specific SDK. Every solution lets you pick Per call or Per cluster so billing matches how you operate.
Federated Learning Aggregator
Replace the central FL aggregator with shared computation. Train one model; only the combined update leaves.
FL Aggregator
Private Queries
Analytics
Key Generation
Per call
By Round (recommended)
Pay per aggregation round.
Scale up to [N parties] and gradient size up to [X size]. Or choose MCU pricing for maximum flexibility
Per cluster
Predictable monthly usage
FL Cluster tiers include a monthly allowance of rounds (e.g., 2k / 10k / 50k) with reserved training-time concurrency.
All tiers include:
Starter (2k rounds/mo)
Flower-friendly adapter SDK
Round monitoring
Solutions Pricing
Use this when you want a well-defined use case that ships faster: each solution includes Lang modules, VM presets, and an app-specific SDK. Every solution lets you pick Per call or Per cluster so billing matches how you operate.
Federated Learning Aggregator
Replace the central FL aggregator with shared computation. Train one model; only the combined update leaves.
FL Aggregator
Private Queries
Analytics
Key Generation
Per call
By Round (recommended)
Pay per aggregation round.
Scale up to [N parties] and gradient size up to [X size]. Or choose MCU pricing for maximum flexibility
Per cluster
Predictable monthly usage
FL Cluster tiers include a monthly allowance of rounds (e.g., 2k / 10k / 50k) with reserved training-time concurrency.
All tiers include:
Starter (2k rounds/mo)
Flower-friendly adapter SDK
Round monitoring
What count's towards usage?
General compute: MCUs = seconds × parties.
Federated learning:
Rounds (or MCUs).
Private queries:
Private ops (or private queries).
Time-series:
Job runs (or windows).
ADKG:
Ceremonies (or participants).
What count's towards usage?
General compute: MCUs = seconds × parties.
Federated learning:
Rounds (or MCUs).
Private queries:
Private ops (or private queries).
Time-series:
Job runs (or windows).
ADKG:
Ceremonies (or participants).
What count's towards usage?
General compute: MCUs = seconds × parties.
Federated learning:
Rounds (or MCUs).
Private queries:
Private ops (or private queries).
Time-series:
Job runs (or windows).
ADKG:
Ceremonies (or participants).
Common Questions
Can I start Per call and switch to Per cluster?
How do I estimate MCUs?
Are Solutions available Self-Hosted?
Do I pay twice (general compute + solution units)?
Common Questions
Can I start Per call and switch to Per cluster?
How do I estimate MCUs?
Are Solutions available Self-Hosted?
Do I pay twice (general compute + solution units)?
Common Questions
Can I start Per call and switch to Per cluster?
How do I estimate MCUs?
Are Solutions available Self-Hosted?
Do I pay twice (general compute + solution units)?
Products
© 2025 Stoffel. All rights reserved.
Products
© 2025 Stoffel. All rights reserved.
Products
© 2025 Stoffel. All rights reserved.