Collate OpenMetadata Pricing and OpenMetadata Cloud Cost at $75,000 a Year
Collate's pricing page lists three tiers and no dollar figures. Its AWS Marketplace listing has one: the Premium Package is $75,000 for 12 months, 25 users and 5,000 data assets. Below is what that buys, how fast its lineage refreshes, what counts against the asset cap, and what running open-source OpenMetadata yourself really costs.
Read-only connection. DataTrail never moves or mutates your data.
OpenMetadata has no license fee. Collate, the managed OpenMetadata from the same team, lists its Premium Package on AWS Marketplace at $75,000.00 for a 12-month contract covering 25 users and 5,000 data assets, which is $3,000 per user or $15 per asset per year. Collate also offers a Free Tier (5 users, 500 assets, weekly lineage) and an Enterprise tier (50 users, 10,000 assets, hourly lineage) with no published price. Premium refreshes lineage once a day. Self-hosting the open-source build needs about $490 a month of AWS compute at the documented minimum, plus the engineer who runs it.
Collate pricing on AWS Marketplace is one package at $75,000
The listing is sold by Collate Inc. itself, not a reseller, and it is transactable: an AWS customer can accept the contract and pay on the AWS invoice. That is why it carries a dollar figure when getcollate.io, which routes Premium and Enterprise to "Book a Demo", does not. The listing has exactly one rate card, and it was the same when we first read it on 21 September.
| Field | What the listing says |
|---|---|
| Package | Collate Premium Package |
| What it covers | Full platform access for 25 users + 5,000 data assets |
| Contract | 12 months, one unit |
| List price | $75,000.00 |
| Other durations offered | None, 12 months only |
| Seller | Collate Inc., the company behind OpenMetadata |
$75,000 divided by 25 users is $3,000 a year, or $250 a month, per user at the cap. The package is one unit, so 10 users pay the same list price and the effective rate climbs to $7,500 each.
Divided the other way, $75,000 over 5,000 assets is $15 per asset. Unlike DataHub's package, Collate meters both people and data, so whichever you run out of first sets your renewal.
Collate's three tiers, and the refresh rate each one buys
These limits come straight from the comparison table on getcollate.io/pricing. The row buyers skip is lineage update frequency, and it is the one that decides whether the graph can answer "what broke this morning".
| Limit | Free Tier | Premium | Enterprise |
|---|---|---|---|
| Users included | 5 | 25 | 50, unlimited on request |
| Data assets included | 500 | 5,000 | 10,000, unlimited on request |
| Lineage update frequency | Weekly | Daily | Hourly |
| Metadata refresh frequency | Weekly | Every 8 hours | Hourly |
| Data quality test cases | Limited to 10 | Unlimited | Unlimited |
| Alerts and notifications | Limited to 5 | Unlimited | Unlimited |
| Single sign-on | No | Yes | Yes |
| Deployment | Multi-tenant SaaS | Single-tenant or hybrid SaaS | Adds private BYOC |
| Published price | No charge | $75,000 a year (AWS Marketplace) | Not published |
Premium updates lineage once a day. A column renamed at 9am shows up in tomorrow's graph. Hourly lineage is an Enterprise feature, and Enterprise has no public price.
Weekly lineage, 10 test cases, 5 alerts, no SSO, and Collate reclaims clusters after four weeks without a login. It shows you the product. It will not run production monitoring.
Collate says you can buy extra users and assets on any tier, and lists AWS VPN and a development cluster as add-ons. None of those prices is published anywhere we could find.
What counts as a data asset, and when 5,000 runs out
Collate's pricing FAQ defines an asset as a table, view, stream, topic, dashboard, pipeline, job, ML model, report or container. Charts, glossary terms, tags, test cases, saved queries, tasks, policies, teams and services are free. So the cap counts your BI layer and your orchestrator, not just the warehouse.
| Illustrative estate | Assets counted | Premium (5,000) |
|---|---|---|
| 800 tables and views, 1,500 dashboards and reports, 400 pipelines and jobs | 2,700 | Fits, with room |
| 2,000 tables and views, 2,500 dashboards and reports, 900 pipelines and jobs | 5,400 | Over the cap |
| 4,000 tables and views, 800 dashboards and reports, 300 topics and ML models | 5,100 | Over the cap |
The second row is the common one. A mid-size analytics team with 2,000 warehouse objects looks comfortably under 5,000 until Tableau, Looker or Power BI is connected and every workbook, dashboard and report starts counting. Add Airflow or dbt Cloud jobs and the pipeline layer adds hundreds more. Count your dashboards before you count your tables.
Two cheap ways to size it before a sales call: run SELECT COUNT(*) FROM information_schema.tables per database for the warehouse side, and export the object inventory from your BI tool's admin console for the rest. If the total is near 5,000, ask Collate in writing for the per-asset add-on price and the Enterprise step-up price, because a successful rollout grows the catalog every month.
What self-hosted OpenMetadata costs to run
The OpenMetadata docs publish a production minimum for each dependency. Priced as plain EC2 on-demand instances in US East (N. Virginia), at 730 hours a month:
| Component | Documented minimum | Closest EC2 size | On-demand per month |
|---|---|---|---|
| Database (MySQL or PostgreSQL) | 4 vCPU, 16 GiB | m6i.xlarge | $140.16 |
| Search (Elasticsearch or OpenSearch), 1 master + 2 workers | 3 x 2 vCPU, 8 GiB | 3 x m6i.large | $210.24 |
| Ingestion (Apache Airflow) | 4 vCPU, 16 GiB | m6i.xlarge | $140.16 |
| Total, before the OpenMetadata server, storage and backups | 14 vCPU, 56 GiB | $490.56 a month, about $5,900 a year |
That is the floor, not the bill. It leaves out the OpenMetadata server itself, 500 GB or more of storage across the three components, backups, and the premium you pay for managed RDS and OpenSearch Service instead of raw instances. It also assumes supported versions: MySQL 8.0.42 or PostgreSQL 15 and up, Elasticsearch 9.x or OpenSearch 3.x, and Airflow 2.10.5.
The larger cost is people. The project moves fast, which is good news for features and a real maintenance load: 2.0.1 shipped on 2 September 2026, 1.13.5 on 3 September, 1.13.6 on 11 September and 2.0.2 on 16 September, four releases across two supported lines in two weeks. If a platform engineer spends a fifth of their time on upgrades, search reindexing and ingestion failures, that salary share is the number to put next to Collate's $75,000.
And check the license before you plan anything commercial on top of it. GitHub reports OpenMetadata as Apache 2.0 because the root LICENSE file is Apache, but the ingestion framework and the UI are under the Collate Community License from release 1.6.0 onward. Internal use is fine. Offering OpenMetadata to others as a hosted service is the excluded purpose. Our OpenMetadata vs DataHub comparison walks through that split file by file.
Collate pricing vs DataHub Cloud, Alation, Collibra and Atlan
None of these catalog vendors prints a price on its own website. Four of the five print one on AWS Marketplace. These are 12-month list prices read from each vendor's own listing.
| Product | 12-month list | What the package covers | Per user | Source, read |
|---|---|---|---|---|
| Collate (managed OpenMetadata) | $75,000 | 25 users and 5,000 data assets | $3,000 per user | AWS Marketplace, 26 Sep 2026 |
| DataHub Cloud | $75,000 | 20 monthly active users | $3,750 per MAU | AWS Marketplace, 21 Sep 2026 |
| Alation | $60,000 | 1 unit, never defined | Not computable | AWS Marketplace, 26 Aug 2026 |
| Collibra | $170,000 | 1 platform unit | Not computable | AWS Marketplace, 26 Aug 2026 |
| Atlan | No price | Private offer only | Not computable | AWS Marketplace, 26 Aug 2026 |
| DataTrail | $708 to $5,748 | Flat plan, no user or asset meter | Flat | datatrail.ai/pricing |
Collate and DataHub Cloud both list at exactly $75,000. Collate gives more seats and caps assets; DataHub gives fewer seats and no data cap. Full breakdown on DataHub pricing.
$60,000 for one unit of "Alation Data Catalog", a unit the listing never defines. Its US federal price list is itemized; see Alation pricing.
$170,000 a year at list, with no discount for a longer term. Atlan shows no price on AWS at all. See Collibra pricing and Atlan pricing.
When a lineage tool beats a $75,000 catalog
Collate is a strong buy for a particular company: one that wants a catalog the business actually opens, with a glossary, ownership, PII classification and AI agents that write documentation, and that would rather pay than run Elasticsearch and Airflow. For that buyer, a fully managed OpenMetadata at $75,000 is competitive with DataHub Cloud and well under Collibra.
Many people pricing Collate are a different buyer. They are a data team that needs to know which dashboards a column feeds before they drop it, and why a model went stale overnight. That is a lineage and monitoring job, it does not need 25 catalog seats, and a graph that refreshes once a day is too slow for it.
DataTrail connects to Snowflake, Databricks, BigQuery, Redshift or Postgres with a read-only role and parses query history and dbt artifacts into column-level lineage. Open any column and impact analysis lists every model, exposure and dashboard downstream before the change merges. Freshness monitoring and schema change alerts run on the same graph. Plans are $59, $179 or $479 a month billed annually on the pricing page, and Team and Scale include unlimited team members, so there is no seat or asset count to forecast.
It is a narrower product. There is no business glossary, no stewardship workflow and no AI documentation agent, and if those are the mandate, Collate or another governance catalog is the right purchase. The honest side-by-side is on OpenMetadata alternatives, and the wider field is in data catalog tools.
Collate and OpenMetadata pricing questions buyers ask
How much does Collate cost?
Collate, the managed version of OpenMetadata, lists its Premium Package on AWS Marketplace at $75,000.00 for 12 months, covering 25 users and 5,000 data assets. That is $3,000 per user or $15 per data asset per year. Collate also runs a Free Tier (5 users, 500 assets) and an Enterprise tier with no published price. getcollate.io itself prints no dollar figures.
What is Collate OpenMetadata pricing?
OpenMetadata, the open-source project, has no license fee. Collate, the company that created it, sells the managed product in three tiers: Free (5 users, 500 assets), Premium (25 users, 5,000 assets, $75,000 a year on AWS Marketplace) and Enterprise (50 users, 10,000 assets, quoted by sales). Extra users and assets can be bought on any tier, but the add-on prices are not published.
Is OpenMetadata free?
The software is free to download and self-host. Running it is not free: the production minimum in the OpenMetadata docs is a 4 vCPU database, a three-node search cluster and a 4 vCPU Airflow instance, roughly $490 a month in AWS compute before the OpenMetadata server, storage and the engineer who keeps it upgraded. Collate's hosted Free Tier costs nothing but caps you at 5 users and 500 assets.
Is OpenMetadata open source?
Partly. The OpenMetadata core server is Apache 2.0, which is what GitHub reports. The ingestion framework and the UI are under the Collate Community License: PyPI shows openmetadata-ingestion 1.5.0 as Apache and every release from 1.6.0 to 2.0.2.0 as Collate Community License. Internal use is allowed; offering a competing hosted service is not. Collate's own pricing FAQ confirms the split.
Collate vs OpenMetadata: what is the difference?
OpenMetadata is the open-source metadata platform you run yourself. Collate is the managed service from the same team, running the same core plus features that are not in the open-source build, including its AI agents (AskCollate, AI Studio, documentation and quality agents), automated PII classification on paid tiers, SSO, and an SLA. You pay Collate to own the infrastructure and upgrades.
What counts as a data asset in Collate pricing?
Per Collate's pricing FAQ, data assets are tables, views, streams, topics, dashboards, pipelines, jobs, ML models, reports and containers. Charts, glossaries, tags, test cases, queries, tasks, policies, teams and services do not count. Dashboards and pipelines counting is what surprises buyers: a BI-heavy team can exhaust 5,000 assets with fewer than 2,000 tables.
Does Collate have a free tier?
Yes. The Collate Free Tier includes 5 users and 500 data assets on multi-tenant SaaS, with lineage updated weekly, test cases limited to 10, alerts limited to 5 and no SSO. Collate reclaims free clusters after four weeks of inactivity. It is built for evaluation, not for running production lineage, since a weekly refresh cannot tell you what broke this morning.
How often does Collate update lineage?
It depends on the tier. Collate's comparison table lists lineage update frequency as weekly on Free, daily on Premium and hourly on Enterprise, and metadata refresh as weekly, every 8 hours and hourly. So the $75,000 Premium package gives you a lineage graph that is up to a day old. Hourly lineage is an Enterprise feature with no published price.
Is Collate cheaper than DataHub Cloud?
At list they cost the same: both sell a $75,000, 12-month package on AWS Marketplace. Collate covers 25 users and 5,000 data assets, so $3,000 per user; DataHub Cloud covers 20 monthly active users with no asset cap, so $3,750 per user. Collate is cheaper if your limit is people, DataHub if your limit is the number of tables and dashboards.
What is a cheaper alternative to Collate for data lineage?
If the job is column-level lineage and knowing what a change breaks, a lineage-first tool on a flat plan costs a fraction of a catalog contract. DataTrail connects read-only to Snowflake, Databricks, BigQuery, Redshift or Postgres and includes column-level lineage, impact analysis, freshness monitoring and schema-change alerts for $59 to $479 a month billed annually, with no user or asset meter.
Where these numbers came from
AWS Marketplace pages render in the browser, so a plain fetch returns nothing useful. We rendered Collate's listing with headless Chrome on 26 September 2026 and read the contract rate card from the page data:
MP=https://aws.amazon.com/marketplace/pp google-chrome --headless=new --dump-dom \ --virtual-time-budget=30000 \ "$MP/prodview-ynw6gz7pzjxm4" # Collate Premium Package # Full platform access for 25 users # + 5,000 data assets # P12M 75000.00
The tier limits, refresh frequencies, the asset definition and the free-cluster inactivity rule come from getcollate.io/pricing on the same day. The hardware minimums come from the OpenMetadata 2.0 deployment docs, EC2 prices are US East on-demand list rates, and the release dates and license fields come from the GitHub and PyPI APIs. DataHub, Alation, Collibra and Atlan figures are this site's own readings of their AWS listings on the dates shown.
What we did not do: we did not repeat per-seat estimates from aggregator sites, because Collate has not published them. We did not quote the AI-generated summary AWS shows on the listing. The asset estates in the sizing table are illustrations, not customer data. Every figure is list, before any negotiated discount or private offer.
Related pages
What OpenMetadata does well, what it takes to run, and the realistic shortlist.
Both managed clouds list at $75,000. Licensing, lineage and which to pick.
DataHub Cloud at $75,000 for 20 monthly active users, and what Core leaves out.
OpenMetadata, DataHub, Amundsen and Atlas, checked for real maintenance activity.
$170,000 a year at list, and why a longer term buys no discount.
The whole field compared on lineage depth, deployment and price.
See your column-level lineage before you sign a catalog contract
Connect your warehouse with a read-only role and get a map of what feeds what, down to the dashboards, with freshness and schema alerts on the same graph. No Airflow to run, no asset cap, a flat monthly plan.