# Estuary

> Estuary is a managed data pipeline platform that moves data from databases, SaaS apps, files and event streams into warehouses, lakes, operational databases and AI systems, using log-based change data capture and streaming with sub-100ms latency or scheduled batch loads, priced at $0.50 per GB moved plus $100 per connector instance per month for the first six ($50 after), with a permanent free tier of 10 GB a month on two connectors and a 30-day free trial without a credit card.

- Category: Customer Data Platforms (https://saastracker.org/categories/customer-data-platforms)
- Website: https://estuary.dev
- Starting price: $0 (Developer, 10 GB a month on 2 connector instances); Cloud $0.50 per GB plus $100 per connector instance per month, billed monthly
- Free plan: Developer: free indefinitely for up to 10 GB a month and 2 connector instances
- Free trial: 30 days of the Cloud plan, no credit card; starts automatically when free-tier limits are exceeded
- Founded: 2019, HQ: New York City and Columbus, Ohio, United States (the vendor lists offices in both and does not name a headquarters), Ownership: Venture-backed
- Profile last reviewed: 2026-09-27
- Canonical profile: https://saastracker.org/products/estuary

## Key facts

| Attribute | Value | Source |
| --- | --- | --- |
| Starting price | $0 (Developer); Cloud $0.50/GB + $100 per connector instance per month, billed monthly | estuary.dev/pricing, 2026-09-27 |
| Free tier volume | 10 GB a month on 2 connector instances, indefinitely | estuary.dev/pricing, 2026-09-27 |
| First paid tier | Cloud: $0.50 per GB plus $100 per connector for the first six, $50 each after; no annual price published | estuary.dev/pricing, 2026-09-27 |
| Pricing meter | GB of data moved plus active connector instances | estuary.dev/pricing, 2026-09-27 |
| Free trial | 30 days of Cloud, no credit card | estuary.dev/pricing, 2026-09-27 |
| Self-hosting | No self-serve edition; private and BYOC deployments on Enterprise annual contracts; code under BSL | estuary.dev/pricing, 2026-09-27 |
| Compliance | SOC 2 Type II, HIPAA | estuary.dev/pricing, 2026-09-27 |

## Overview

Estuary was founded in 2019 by Dave Yaffe (CEO) and Johnny Graettinger (CTO), according to the vendor's about page, to build on Gazette, a streaming framework the team had written inside the ad-tech company Arbor in 2014 and run at LiveRamp after LiveRamp acquired Arbor in 2016. FirstMark invested $7 million in 2021, the self-service cloud product launched in 2023, and a $17 million Series A led by M13 followed in 2025. The company lists offices in New York City and Columbus, Ohio.

The product's argument is that one pipeline tool should handle both speeds. Database sources such as Postgres, MySQL, MongoDB, Oracle and SQL Server are read through log-based change data capture after an initial backfill, SaaS sources such as HubSpot, Salesforce, Stripe, Shopify, Klaviyo and the major ad platforms are pulled incrementally or by webhook, and every change lands in a collection stored in cloud storage the customer can own. Each destination then gets its own cadence, from sub-second streaming to an hourly batch, so a team can feed a live dashboard and a cost-sensitive warehouse from the same capture without extracting twice.

Pricing is published and simple to model. The Developer tier is free indefinitely for up to 10 GB a month on two connector instances. Beyond that the Cloud plan charges $0.50 per GB of data moved plus $100 a month for each of the first six connector instances and $50 for each one after, billed monthly by card or 30-day invoice. The page has no monthly and annual toggle: annual commitments exist as pre-paid data packages usable over up to 12 months, with discounts the vendor describes but does not publish, and private or bring-your-own-cloud deployments are annual Enterprise contracts.

For a go-to-market team this is warehouse plumbing rather than a marketer's CDP. Estuary does not collect website or app events with a tracking snippet and does not build customer profiles or audiences; it moves the data that already exists in the CRM, billing system, ad accounts and production database into the place where those profiles get built. The buyer is usually the first data or analytics engineer at a startup, and the startup program ($2,000 in credits over 12 months for companies that have raised less than $5 million) is aimed squarely at that person.

## How it works

1. You create a free account without a card, pick a source connector (a database, a SaaS app, a file store or a stream such as Kafka or Kinesis) and enter credentials, optionally through an SSH tunnel for a firewalled database. Estuary runs a historical backfill and then switches to log-based CDC or incremental pulls.

2. Captured data is written to collections, durable logs stored in Estuary's storage during a trial or in a cloud bucket you configure, which lets new destinations be added later without re-extracting from the source.

3. Materializations write collections to destinations such as Snowflake, BigQuery, Redshift, Databricks, Postgres, Elastic, MongoDB or Pinecone, each on its own schedule from real-time streaming to hour-plus intervals. Transformations, called derivations, run in SQL, Python or TypeScript in flight, and dbt Cloud jobs can be triggered when new data lands.

4. Pipelines can be built in the web UI or as configuration with the flowctl CLI, and Agent Skills let coding assistants such as Claude Code or Cursor draft pipeline configuration for a human to review before it is published. Once free-tier limits are exceeded, a 30-day trial of the Cloud plan starts automatically, after which usage is billed monthly.

## Best for

Startups and small data teams that need production database changes and SaaS data (CRM, billing, ads) in a warehouse or operational system within seconds or minutes, and want a published per-GB price instead of row-based billing.

## Not the right fit for

- Marketing teams looking for a CDP that collects website events, stitches identities and builds audiences. Estuary moves data between systems; it does not track visitors or manage profiles.
- A business without a warehouse, database or engineer to own the destination. The output is tables, not reports.
- Buyers who want a fixed annual price on the page. Annual pre-pay discounts and Enterprise pricing are negotiated.
- Teams that must self-host an open-source pipeline. The code is published under the Business Source License, and private or bring-your-own-cloud deployments are Enterprise contracts.
- Very small monthly volumes across many sources. Connector fees of $100 each for the first six can outweigh the data charge when many low-volume SaaS tools are connected.

## Features

### Capture

Where Estuary differs most from batch ELT tools.

- **Log-based change data capture**: Streams inserts, updates and deletes from database logs rather than querying tables, for Postgres, MySQL, MongoDB, Oracle and SQL Server among others, which the vendor says minimizes load on production systems.
- **Historical backfill then CDC**: Each pipeline backfills existing data first and then switches to real-time capture; backfills can be scoped or disabled for large tables.
- **200-plus managed connectors**: Databases, cloud storage, SaaS apps and streams, including HubSpot, Salesforce, Stripe, Shopify, Klaviyo, Iterable, Braze, Intercom, Mixpanel, Amplitude, Google Ads, Meta Ads, LinkedIn Ads and TikTok. The startup page gives 150-plus sources.
- **Webhook, file and HTTP ingestion**: Generic webhook, CSV and JSON file drop, and custom HTTP connectors for sources without a dedicated connector.
- **Kafka and streaming sources**: Captures from Kafka, Kinesis and Google Pub/Sub, and the plan includes Kafka compatibility for downstream consumers.

### Storage and delivery

Capture once, deliver to many places at different speeds.

- **Collections in your own storage**: Captured data is stored as durable collections that can live in your own cloud bucket for as long as you set, so the source is extracted only once.
- **Per-destination latency**: Each destination can run from sub-second streaming to hour-plus intervals, balancing freshness against warehouse compute costs.
- **Exactly-once delivery**: The vendor states exactly-once semantics with checkpointing, so retries do not create duplicates.
- **One-to-many materialization**: One source can feed a warehouse, a search index, an operational database and a vector store at the same time.
- **Snowpipe Streaming and AI destinations**: Supports Snowflake's Snowpipe Streaming, plus Databricks, BigQuery, Redshift, Elastic, MongoDB, DynamoDB and Pinecone.
- **Replay and time travel**: Collections can be replayed to backfill a new destination or reprocess history without touching the source.

### Transformation and schema

Enough to shape data in flight without a separate stream processor.

- **Derivations in SQL, Python or TypeScript**: Filter, join, aggregate and enrich data in motion; simple renames and field removal can be done in the dashboard without code.
- **dbt Cloud integration**: Triggers dbt Cloud jobs automatically when new data arrives in the destination.
- **Schema inference and evolution**: Detects new fields, changed types and new streams automatically, or lets you approve changes manually.
- **Preview and test**: Derivations can be previewed and validated against sample data in the UI or CLI before publishing.

### Operations and deployment

What a small team needs to run it, and what Enterprise adds.

- **Web UI and flowctl CLI**: Pipelines can be built without code in the dashboard or managed as configuration from the command line.
- **Agent Skills and MCP**: Open-source skills let coding assistants draft, publish and troubleshoot pipelines with human review before deployment, and an MCP integration answers questions about the platform.
- **Monitoring and OpenMetrics**: Pipeline status, latency and metrics in the dashboard, with an OpenMetrics API for Prometheus, Datadog and similar tools, plus alerting on the Cloud plan.
- **US and EU processing regions**: Public data planes in US Central (Google Cloud), US East, US West, Europe (Ireland) and Asia Pacific (Sydney) on AWS, per the security page.
- **Private and BYOC deployments**: Enterprise customers can run the data plane in their own VPC or cloud, with PrivateLink and Private Service Connect, on annual contracts.

## Use cases

- **First data engineer at a Series A SaaS company**: Product usage lives in Postgres, revenue in Stripe and pipeline in HubSpot, and the nightly batch job means dashboards are always a day behind. Outcome: Postgres CDC, Stripe and HubSpot captures and a Snowflake materialization run as four connector instances, billed at $100 each plus $0.50 per GB, with the company possibly qualifying for $2,000 in startup credits.
- **Growth lead at an ecommerce brand**: Shopify orders, Klaviyo engagement and ad spend from Google and Meta sit in four tools, and nobody can see blended acquisition cost until month end. Outcome: The sources land in BigQuery on an hourly cadence to keep warehouse costs down, while an order stream feeds an operational database in near real time from the same captures.
- **Analytics engineer replacing a batch ELT tool**: Row-based billing climbs every time a high-churn table is added, and the team cannot forecast the invoice. Outcome: Estuary bills by data volume and connector count, so the invoice can be modeled from GB moved, and the backfill lets the team cut over table by table.

## Pricing

Usage-based: a charge per GB of data moved plus a flat monthly fee per active connector instance (each source or destination), with a permanent free tier, a 30-day Cloud trial, optional annual pre-pay packages at unpublished discounts, and quoted Enterprise contracts.

- **Developer**: $0 free indefinitely. Up to 10 GB a month; 2 concurrent connector instances; No credit card or billing details.
- **Cloud**: $0.50 per GB + $100 per connector per month, billed monthly; connector instances after the sixth are $50 each. 200-plus managed connectors; Unlimited users; Real-time or batch at no extra charge; RBAC, SSH tunnels, bring your own storage; US and EU processing regions; dbt Cloud integration and Kafka compatibility; Community Slack and email support. Paid by card, debit card or 30-day invoice. No annual price is shown on the page.
- **Enterprise**: Custom quoted; private and BYOC deployments on annual contracts only. Volume-based discounts; SOC 2 and HIPAA compliance reports; Single sign-on; Custom SLA terms; Private and BYOC deployments; Dedicated Slack support and migration assistance.

Add-ons:

- Startup program ($2,000 in credits over 12 months): For companies that have raised less than $5 million and are not current or former customers.

Billing notes:

- There is no monthly and annual toggle on the pricing page. Pay-as-you-go is billed monthly; pre-pay buys a fixed data amount usable over up to 12 months, with larger discounts for larger payments, but the discount rates are not published.
- A connector instance is one source capture or one destination materialization, so a single source-to-warehouse pipeline is two instances. The first six cost $100 a month each and each additional one $50, prorated.
- The billing FAQ says data usage covers volume sourced, transformed and delivered, while the pricing examples say each source and target is billed once; model a transformation-heavy pipeline with the calculator before committing.
- Exceeding the free tier starts the 30-day trial automatically; after it, the account either moves to usage billing or pipelines are paused, and the vendor says it warns before billing.
- Private and bring-your-own-cloud deployments, SSO and compliance reports are Enterprise-only.

Value assessment: Estuary's pricing is easier to forecast than row-based ELT: a pipeline from one database and two SaaS apps into one warehouse is four connector instances, $400 a month in connector fees, plus fifty cents per GB moved, and the free tier covers a real first pipeline at small volume. The model favors high-volume, high-change sources such as production databases, where row-based billing grows fastest, and penalizes many small SaaS sources, where the $100 connector fee dominates. Real-time delivery costs nothing extra, which is the main economic difference from tools that charge for sync frequency. The weak spot for a small team is the floor once past the free tier: even a modest stack runs to a few hundred dollars a month before data volume.

### Entry price history

Dates are when SaaSTracker checked the vendor's published pricing, not when the vendor changed it.

Entry price first recorded on 27 September 2026. It will be re-checked when this category next comes round the rotation.

| Checked | Entry price | Normalised | Event | Note |
| --- | --- | --- | --- | --- |
| 2026-09-27 | $0 (Developer, 10 GB a month on 2 connector instances); Cloud $0.50 per GB plus $100 per connector instance per month, billed monthly | $10 monthly, flat | recorded |  |

## Strengths

- Published usage pricing with a permanent free tier of 10 GB a month on two connectors and a 30-day trial without a card.
- Log-based change data capture with sub-100ms end-to-end latency, and batch on the same platform with per-destination scheduling.
- Capture once, deliver many: collections stored in your own bucket let new destinations and replays run without re-extracting the source.
- Transformations in SQL, Python or TypeScript in flight, plus dbt Cloud triggers after load.
- SOC 2 Type II and HIPAA, US and EU processing regions, SSH tunnels and bring-your-own storage on the self-serve plan.
- A startup program worth $2,000 in credits for companies that have raised under $5 million.

## Limitations

- Not a behavioral CDP: no website or app tracking snippet, identity resolution or audience builder.
- Connector fees make many low-volume SaaS sources relatively expensive.
- No annual price or pre-pay discount is published, and Enterprise features such as SSO require a quote.
- Source code is under the Business Source License, and there is no self-serve self-hosted edition.
- Connector count is given as 200-plus on the pricing page and 150-plus sources on the startup page, so check the specific connectors you need.

## Comparisons

- **Estuary vs Fivetran**: Fivetran is the reference managed ELT service, billed on monthly active rows and built around batch syncs. Estuary bills by GB moved plus connectors, streams database changes in real time at no extra charge, and stores captures so new destinations do not re-extract. Choose Fivetran for the broadest catalogue of mature SaaS connectors; choose Estuary when database CDC, low latency or predictable per-GB billing matter more.
- **Estuary vs Airbyte**: Airbyte is open source with a large connector catalogue and a connector-building kit, and can be self-hosted for free. Estuary is managed-first, source-available rather than open source, and built around streaming CDC with exactly-once delivery. Teams that want to run their own pipelines or build obscure connectors should start with Airbyte; teams that want real-time replication without operating infrastructure should look at Estuary.
- **Estuary vs Hevo Data**: Hevo publishes an event-based ladder, free to 1 million events and $299 a month ($265 annually) for 5 million on Starter, with no-code pipelines into warehouses. Estuary's free tier is sized in GB and its paid meter in GB plus connectors, and it adds sub-second streaming and one-to-many delivery to operational systems. For a handful of SaaS sources into one warehouse, Hevo's fixed tiers are simpler to budget; for database CDC feeding several destinations, Estuary fits better.

## Pairs well with

Estuary is a CDP tool, and a team running it still has 7 neighbouring categories to staff. The pick in each is Mixpanel for product analytics & session replay, Umami for GTM analytics, EngageBay for marketing automation, Sender for email service providers, ChartMogul for retention & churn prevention, Make for GTM engineering and Textedly for SMS marketing, each the leading tool in that category on published criteria.

| Category | Ideal pairing | Because | Benefit |
| --- | --- | --- | --- |
| Product Analytics & Session Replay | [Mixpanel](https://saastracker.org/products/mixpanel) (Best Value) | Estuary is built for CDP, not product analytics. Mixpanel adds a free tier of 1 million events and 10,000 replays per month. | Feature usage and drop-off are measurable, so a roadmap argument can end with evidence. |
| GTM Analytics | [Umami](https://saastracker.org/products/umami) (Best Value) | Estuary lists not a behavioral CDP among its limits. Umami adds a free Cloud tier at roughly 100,000 events a month with three. | Spend gets attributed to revenue, so the channel that works is the one that gets funded. |
| Marketing Automation | [EngageBay](https://saastracker.org/products/engagebay) (Best Value) | Estuary is built for CDP, not automation. EngageBay adds bringing your own sending provider. | Follow-up fires off behaviour rather than a calendar, so nobody is chased at the wrong moment. |
| Email Service Providers | [Sender](https://saastracker.org/products/sender) (Best Value) | Estuary lists connector fees make many low-volume SaaS sources relatively expensive among its limits. Sender adds SMS and email combined inside a single automation flow rather. | An opt-in audience is owned rather than rented, and reaching it costs nothing per send. |
| Retention & Churn Prevention | [ChartMogul](https://saastracker.org/products/chartmogul) (Best Value) | Estuary is built for CDP, not retention. ChartMogul adds a genuinely free plan up to $10,000 MRR with full metrics. | Accounts at risk surface while there is still time to call, rather than at renewal. |
| GTM Engineering | [Make](https://saastracker.org/products/make) (Best Value) | Estuary is built for CDP, not GTM engineering. Make adds best power-per-dollar in mainstream automation. | The workflow nobody wanted to buy a whole tool for gets built in an afternoon. |
| SMS Marketing | [Textedly](https://saastracker.org/products/textedly) (Best Value) | Estuary is built for CDP, not SMS. Textedly adds every feature on every plan. | The urgent message gets read within minutes, on the one channel almost nobody ignores. |

Pairings follow the SaaSTracker methodology (https://saastracker.org/methodology): the leading tool in each complementary category, chosen on published criteria. Placement cannot be purchased.

## Implementation

- Setup time: Minutes for a first pipeline in the UI, according to the vendor; a production database CDC setup typically needs database log settings, a replication user and possibly an SSH tunnel, which is closer to a day of engineering work.
- Learning curve: Low for connecting sources and destinations in the UI; moderate for derivations, schema evolution policies and flowctl configuration-as-code.
- Onboarding: Self-serve with documentation, tutorials and a public Slack community with an AI bot. Enterprise adds dedicated Slack support and migration assistance.
- Migration: Run Estuary alongside the existing ELT tool, backfill each table, and cut over destination by destination. Keep collections in your own bucket from the start so later destinations can be added without another backfill.

## Platform, API & security

- Platforms: Web app, flowctl CLI, Agent Skills for AI coding assistants
- API: flowctl CLI for configuration-as-code, an OpenMetrics API for monitoring, API references in the documentation, and an MCP integration.
- Compliance: SOC 2 Type II, HIPAA
- Data residency: US and EU processing regions on the Cloud plan; public data planes in US Central, US East, US West, Europe (Ireland) and Asia Pacific (Sydney); private data planes in any AWS, Google Cloud or Azure region on Enterprise.
- SSO: Enterprise only.
- Security notes: TLS for all traffic including mutual TLS between internal services, AES-256 at rest, RBAC with time-limited tokens, SSH tunnels on Cloud, and PrivateLink or Private Service Connect for private deployments.

## Support

- Channels: Public Slack community, Email (support@estuary.dev), Dedicated Slack channel on Enterprise, Sales engineer calls, GitHub issues
- Documentation: docs.estuary.dev, with setup guides, tutorials, API references and an AI assistant, plus a public status page.
- Community: Public Slack server with an #ask-ai bot.

## Company

- Founded: 2019
- Founders: Dave Yaffe, Johnny Graettinger
- Headquarters: New York City and Columbus, Ohio, United States (the vendor lists offices in both and does not name a headquarters)
- Ownership: Venture-backed
- Funding: $7 million from FirstMark in 2021 and a $17 million Series A led by M13 in 2025, per the vendor's about page. The founding year of 2019 is stated on the same page.

Funding history:

- Early investment (2021): $7M. From FirstMark; the vendor does not name the round.
- Series A (2025): $17M. Led by M13.

Timeline:

- 2014: Gazette, the streaming framework Estuary is built on, is created inside the ad-tech company Arbor.
- 2016: LiveRamp acquires Arbor; Gazette goes on to handle billions of daily ad-tech events there.
- 2019: Estuary is founded by Dave Yaffe and Johnny Graettinger.
- 2021: FirstMark invests $7 million.
- 2023: Launches its self-service SaaS product for building real-time pipelines.
- 2025: Raises a $17 million Series A led by M13.

## Integrations

PostgreSQL, MySQL, MongoDB, SQL Server, Oracle, Snowflake, Google BigQuery, Amazon Redshift, Databricks, HubSpot, Salesforce, Stripe, Shopify, Klaviyo, Google Ads, Meta Ads, LinkedIn Ads, Apache Kafka, dbt Cloud, Pinecone

## FAQ

### How much does Estuary cost?

The Developer tier is free for up to 10 GB a month on two connector instances. The Cloud plan charges $0.50 per GB moved plus $100 a month for each of the first six connector instances and $50 for each additional one, billed monthly. Enterprise is quoted.

### Is there an annual plan or discount?

Not on the pricing page. Estuary offers pre-paid data packages usable over up to 12 months, with larger discounts for larger payments, and volume discounts on Enterprise, but the rates are not published.

### Is there a free trial?

Yes. A 30-day trial of the Cloud plan starts automatically when you exceed the free tier, and it needs no credit card.

### What counts as a connector instance?

Each source capture and each destination materialization is one instance. A pipeline from one database to one warehouse uses two.

### Is Estuary a customer data platform?

Not in the behavioral sense. It does not collect website events or build customer profiles. It moves data from databases, SaaS tools and streams into warehouses and operational systems, where profiles and audiences are built.

### Does Estuary support real-time and batch?

Both. Databases are captured with log-based CDC, and each destination can be set from sub-second streaming to hour-plus intervals, with no extra charge for real-time.

### Can I transform data in Estuary?

Yes. Derivations transform data in flight in SQL, Python or TypeScript, simple renames and removals can be done in the dashboard, and dbt Cloud jobs can be triggered after data lands.

### Is Estuary open source or self-hostable?

The code is on GitHub under the Business Source License, with some directories also under Apache 2.0. There is no self-serve self-hosted plan; private and bring-your-own-cloud deployments are Enterprise contracts.

### Is Estuary SOC 2 and HIPAA compliant?

The vendor states SOC 2 Type II certification and HIPAA compliance. The compliance reports themselves are listed as an Enterprise feature.

### Does Estuary have a startup program?

Yes. Companies that have raised less than $5 million and are not current or former customers can apply for $2,000 in credits valid for 12 months.

### What should I pair Estuary with?

Estuary is a CDP tool, and a team running it still has 7 neighbouring categories to staff. The pick in each is Mixpanel for product analytics & session replay, Umami for GTM analytics, EngageBay for marketing automation, Sender for email service providers, ChartMogul for retention & churn prevention, Make for GTM engineering and Textedly for SMS marketing, each the leading tool in that category on published criteria.

### What do I pair with Estuary for Product Analytics & Session Replay?

Mixpanel. Estuary is built for CDP, not product analytics. Mixpanel adds a free tier of 1 million events and 10,000 replays per month.

### Is Estuary enough on its own?

For CDP, yes. Estuary does not cover GTM analytics, marketing automation and email service providers, and its own profile says so; Umami, EngageBay and Sender are the tools that fill those gaps.

## Editorial verdict

Estuary is the pipeline tool for a small data team that wants database changes and SaaS data delivered in seconds rather than overnight, at a price it can model. The meter is published and simple, fifty cents per GB plus a flat fee per connector, the free tier runs a real first pipeline indefinitely, and real-time delivery costs nothing extra. It is not a marketer's CDP: there is no tracking snippet, identity graph or audience builder, and its value appears only once there is a warehouse or operational system on the other end. Watch the connector fees if you plan to connect many small SaaS tools, and ask for the pre-pay rates before committing annually, because they are not on the page.

---

Source: SaaSTracker (https://saastracker.org), an independent editorial project. This profile is compiled from public information, carries no peer reviews or paid placement, and was last reviewed 2026-09-27. Awards are judged on published criteria: https://saastracker.org/methodology