Feast
feast.dev
Open-source feature store for machine learning.
Quick Information
- Category:
- MLOps & Infrastructure
- Pricing:
- Free
- Platforms:
- Python, Kubernetes, AWS
- Free Plan:
- Yes
- API Available:
- Yes
- Last Verified:
- October 2026
What Is Feast?
Feast is an open-source feature store for machine learning, licensed under Apache 2.0 and free to use, though you pay for the infrastructure and engineering time to run it. There is no official managed Feast-as-a-service offering, so teams pay for their own cloud resources and operations, while Tecton is the commercial managed alternative on a subscription model. One blog's benchmark, at 10 million users, 200 features and 5,000 lookups a second, puts Feast on AWS (Redis ElastiCache, S3 or Iceberg, and a small EKS for the server) at around $4,200 a month all-in compute, and estimates Tecton starts around $25,000 a month, though that sentence is truncated in the result, and it says DIY on the same infrastructure lands around $3,600 a month with the savings eaten by about 1.5 engineers of operational burden. Treat these as indicative only, as they come from a single source.
Feast is still pre-1.0 but kept up monthly releases in 2026: an Iceberg source arrived in 0.64 and an Iceberg REST catalog and OpenLineage ingestion in 0.65, and another source lists v0.66.0 tagged on August 21, 2026, with v0.64.0 on June 13 and v0.65.0 on July 20. One tracker lists a high-severity advisory from June 2026 (CVE-2026-55563) involving CI integration tests that run untrusted fork code, and gives Feast an OpenSSF Scorecard of 5.5 out of 10, which I could not verify against the official GitHub advisory. Most results are SEO-style comparison blogs, so check Tecton directly for pricing.
Feast suits ML platform teams that want a vendor-neutral feature store and can operate it. Compare Hopsworks, lakeFS and ZenML. Browse the MLOps & Infrastructure tools directory for more.
What Can You Use Feast For?
Feature storage and serving
Online and offline.
Consistent training and inference features
Shared definitions.
Feast Features
- Pre-1.0 monthly releases — v0.66.0 on August 21, 2026.
- Iceberg and OpenLineage — Added in 0.64 and 0.65.
- No managed Feast — Self-run.
Who Is Feast For?
- ML platform teams — Vendor-neutral.
- Real-time ML teams — Online serving.
Feast Pros & Limitations
What We Like
- Free and open source
- Active releases
- Vendor-neutral
Limitations
- You operate it yourself
- Still pre-1.0
- Reported security advisory unverified
Feast Pricing
| Plan | Price | Main Features |
|---|---|---|
| Open source | Free (Apache 2.0) | You pay infrastructure and operations |
| Indicative self-hosted cost | About $4,200/month on AWS for 10M users (single blog benchmark) | Plus about 1.5 engineers of operations |
| Tecton (managed alternative) | Estimated from about $25,000/month (single blog) | Not Feast |
Pricing was last verified October 2026. Check the official website for the latest pricing.
AI Listing Jungle Ranking
- Global Rank:
- #22
- Overall Score:
- 92.5/100
- Monthly Users:
- –
- Rank Change:
- ▲ 1538
This ranking reflects AI Listing Jungle's own independent scoring methodology and is not an endorsement by Feast or its company. How Our Rankings Work
Frequently Asked Questions
Is Feast free?
Yes, Apache 2.0, but you pay for infrastructure.
Is there managed Feast?
No official one; Tecton is a commercial alternative.
Is it production ready?
Used widely but still pre-1.0.
What does it cost to run?
About $4,200 a month in one benchmark.
Our Take
Feast is the open feature store default.
Budget for the ops work.
Feast Alternatives
Feast Comparisons
Ready to try Feast?
Visit the official website to get started.
Visit Official WebsiteAI Listing Jungle is an independent AI tools directory. Product names, logos, and trademarks belong to their respective owners.