Home / Directory / AI SaaS tooling / MLOps & experiment tracking / Union.ai
Union.ai
Union.ai is a Seattle-area company that sells a managed AI runtime built on Flyte, the open-source orchestrator its founders created at Lyft, for running data processing, model training, inference, and agents inside a customer’s own cloud.
ML platform and research engineering teams that run large training, batch inference, or bioinformatics workloads on Kubernetes and want them to recover from failures without reruns.
Small data science teams that mainly want a hosted experiment dashboard, or buyers who want a flat per-seat price.
Verdict
Union.ai grew out of Flyte, the workflow engine its founders built for machine learning at Lyft, and it now sells a managed runtime that schedules data processing, training, inference, and agent runs across clusters and clouds while keeping customer data inside the customer’s own boundary. It is highly recommended for ML platform teams that run heavy, failure-prone compute and want caching, lineage, and recovery built in, and it is more infrastructure than a small team needs to log a few experiments.
Score Breakdown
These scores cover how well a team can log and compare experiments, how well it versions data, models, and other artifacts, how easily it runs work across clouds and clusters, and how well it fits enterprise security and audit needs, and then how the company does on innovation, communication, pricing, and contract terms.
Who Uses Union.ai, Where It Fits on Price, and Its Competitors
Union.ai sells to ML platform, research engineering, and data teams at biotech companies, autonomous systems makers, geospatial firms, and travel and logistics companies. Customers quoted on its site include Woven by Toyota, Hopper, Artera AI, and Prima Mente, and the company says Flyte has more than 1 million monthly downloads. Union.ai was founded in 2021 and raised a $38.1 million Series A led by NEA.
The Team plan carries a $950 monthly minimum that is credited back as usage, and Enterprise contracts are an annual commit, so Union.ai sits in the upper-middle of this field on cost for a production team. It is most often compared with ZenML for open-source pipeline orchestration, with Kubeflow and Ray for teams that run their own clusters, and with ClearML and Valohai for managed MLOps on a team’s own compute.
Pricing
Union.ai publishes its Team plan and usage rate card and quotes Enterprise contracts, as of November 2026.
| Plan | Price | What stands out |
|---|---|---|
| Flyte (open source) | Free | The open-source orchestrator that Union.ai is built on, run on a team’s own Kubernetes clusters. |
| Team | $950 / month minimum, credited to usage | Two active clusters, 10,000 concurrent actions, 30-day data retention, and standard support, with the first $950 of usage free. |
| Enterprise | Custom annual commit | Unlimited clusters across regions and clouds, SSO and SAML, custom roles, a self-hosted option, and 24x7 support. |
The Field at a Glance
How Union.ai compares with other MLOps and experiment tracking vendors we reviewed, by Overall Score and relative typical engagement cost.
Union.ai scores 8.3 in MLOps and experiment tracking, under Weights and Biases (8.9), and ahead of ZenML (8.0), Hopsworks (7.8), ClearML (7.7), lakeFS (7.5), Valohai (7.1), DagsHub (6.8), Comet (6.7), and Domino Data Lab (6.5). Relative cost lands mid-band for this subcategory.
Compared with …
- Union.ai vs ZenML 8.3/8.0
How Well Union.ai Fits Each MLOps Use Case
We rated Union.ai against the jobs ML platform teams most often give an orchestration and MLOps platform, drawing on its site, documentation, customer stories, and pricing page.
| Use case | Fit | Notes |
|---|---|---|
| Large training and fine-tuning runs | Strong | Runs span clusters and GPUs, cache completed steps, and recover from out-of-memory failures without restarting. |
| Batch inference and data processing | Strong | Customers run inference jobs that exceed a single cluster and fan out to thousands of GPUs. |
| Production agents | Strong | Agent steps are checkpointed, and agent-written code runs in its own sandbox on the customer’s nodes. |
| Regulated or sovereign deployments | Strong | Runs in the customer’s cloud, data center, or an air-gapped site, with no customer data passing through the control plane. |
| Lightweight experiment logging | Mixed | Lineage and run tracing are built in, but teams that only want a metrics dashboard often pair it with a dedicated tracker. |
Review Excerpts
These excerpts draw on Union.ai’s site, public customer statements, and public review sites. Paid reviews and pay-for-play sites were excluded.
“Union.ai has been critical at the time that we needed to significantly scale up. Union.ai’s wealth of expertise has enabled us to move fast and deliver at scale.”
“Definitely easier to scale than Ray, since you have a lot more granular control over streaming the parallelism. The parallel async model is nice to work with.”
“Once it was running it saved us a lot of reruns, but the first weeks took real Kubernetes and Flyte know-how, and the usage meters took a while to forecast.”
Representative buyer view
Where customer reviews couldn't be found, our analysts attempted to give a voice to buyer sentiment, drawing on real information from review sites, forums, and product documentation.
Company Facts
- Headquarters
- Seattle area, Washington
- Founded
- 2021
- Funding
- $38.1 million Series A led by NEA
- Pricing model
- Open-source Flyte is free; Team plan has a $950 monthly minimum credited to usage; Enterprise is an annual commit
- Website
- union.ai
Key Leadership
Ketan Umare
Co-founder and CEO
He led the engineering team that built Flyte at Lyft and held engineering roles at Amazon and Oracle.
Methodology
We scored Union.ai on eight criteria, from experiment tracking to contract fairness, and the Overall Score is their simple average on a 0 to 10 scale. The scores draw on reviews from public review sites and forums, with paid reviews and pay-for-play sites excluded, plus a hands-on look at the product, its pricing, and its documentation and training. Scores can change as we add reviews and product checks.
← Back to MLOps & experiment tracking
Sources
Company facts
- Founded: Union.ai about
- Funding: Union.ai Series A
- Headquarters: GeekWire
- Pricing model: Union.ai pricing
Review excerpts
- Excerpts based on: Union.ai and Union.ai pricing
