Perpetual ML
Verified Oct 15, 2026 ActiveML platform native to Snowflake that trains, monitors, and retrains models continually instead of from scratch each cycle.
Listed on EffectHub ·
Pricing
- Pricing model
- Freemium
- Paid options from
- $9.90/request
- Billing
- Metered · per request
Marketplace tasks (auto train, continual learning, monitoring and similar) are free for 1-9 tasks a month, $9.90 per task per month for 10-99 and $0.90 per task at 100+; compute bills at $0.00014 per vCPU-second, storage at $0.000069 per GB-hour and egress at $0.10 per GB.
Use tool ↗About Perpetual ML
Perpetual ML is a machine learning platform built to run inside a Snowflake data warehouse, with Databricks support planned, so data does not need to move elsewhere for training. It covers automated model training, continual learning that updates models from their current state rather than retraining from zero, experiment tracking, a model registry, monitoring, and deployment, and lets users optimize toward custom business objectives rather than accuracy alone. Pricing is usage-based and published openly: compute and storage are billed per vCPU-second and per GB-hour, and a separate marketplace table charges per automated task, free for up to 9 tasks and $0.90/task at volumes over 100. No company name, founding date, or team information is published on the site.
- Platforms
- Web
- API
- Not confirmed
- Last verified
- Oct 15, 2026
Verified sources
- Pricing page: 1-9 tasks free, 10-99 tasks $9.9 per task per month, 100+ tasks $0.9 per task per month; shared CPU $0.00014 per vCPU second, storage $0.000069 per GB hour, egress $0.10 per GB (15 October 2026)perpetual-ml.com ↗
- Home page: auto train with PerpetualBooster, continual learning, experiment tracking, registry, monitoring, deployment, Marimo notebooks; native Snowflake integration with Databricks support upcoming (15 October 2026)perpetual-ml.com ↗
Is Perpetual ML legit?
We don’t label tools safe or unsafe. Below are the records we checked ourselves, each with the date we read it.
- Domain registered
- Feb 5, 2024 (2 years ago)Registry record (RDAP), registrar Cloudflare, Inc., looked up Sep 6, 2026
- TLS certificate
- Issued by Google Trust Services, valid until Jan 11, 2027Seen in our own connection to the site on Oct 15, 2026
- Last checked
- Oct 15, 2026The site answered
Pros & cons of Perpetual ML
Perpetual ML targets teams already living in Snowflake who want continual model retraining and monitoring without exporting data to a separate MLOps platform. The usage-based pricing is unusually transparent for this category, published down to the per-vCPU-second rate, but there's no public information about who runs the company or how long it has operated.
Pros
- Usage-based pricing is published in detail (per-vCPU-second, per-GB-hour, per-task) rather than hidden behind a sales call
- Continual learning avoids full retraining from scratch on every cycle
- Covers the full lifecycle: training, registry, monitoring, and deployment in one product
Cons
- No company name, founding date, or team information published anywhere on the site
- Only Snowflake is supported today; Databricks is listed as upcoming, not available
Best for
- Snowflake-native teams wanting in-warehouse model training and monitoring
- Teams that want continual retraining instead of full retrains on a schedule
Not for
- Teams not already using Snowflake or planning to adopt Databricks
- Buyers who want to vet a company's track record before committing budget
Alternatives to Perpetual ML
Perpetual ML head-to-head
Side-by-side fact sheets against the competitors our research or their makers name — pricing, platforms, API and both verdicts on one page.