What GlassFlow does

GlassFlow GmbH is a Berlin-based, venture-backed startup founded in 2023 by Armend Avdijaj (CEO) and Ashish Bagri (CTO) that builds an open-source, serverless streaming ETL platform purpose-built for ClickHouse ingestion. The platform is positioned as a "next-gen data streaming infrastructure" company targeting Python developers and data engineers who need to build event-driven pipelines without operating Java/Scala-based stacks such as Flink or Kafka Streams. Its core technical differentiator is solving ClickHouse-specific ingestion failure modes — safe historical backfills that do not overwrite live CDC updates, built-in CDC for PostgreSQL (WAL), MySQL (binlog), MongoDB and Debezium, schema evolution handling, configurable deduplication via BadgerDB, ClickHouse-friendly block-size batching, and a Dead-Letter Queue for fault isolation. The platform is Kubernetes-native with Helm-chart deployment, uses NATS JetStream as an internal message broker, supports multi-architecture images (AMD64 + ARM64), and is benchmarked at roughly 510k events/sec peak throughput at 10 replicas.

GlassFlow's commercial offering is built on a hybrid PLG-plus-enterprise motion. The open-source `clickhouse-etl` repository on GitHub serves as the top-of-funnel developer entry point; a free/open-source tier, a quote-based Managed Cloud Service (BYOC, Bring Your Own Cloud), and a quote-based Self-Hosted Enterprise deployment comprise the three main SKUs, with 24/7/365 support, SLAs, and implementation engineers available as add-ons. No pricing is publicly disclosed. Distribution combines self-serve open-source adoption, community-led growth via a Slack community (GlassFlow Hub), developer-focused content marketing around ClickHouse best practices, and direct enterprise sales. Documented customers include Nubrain (AI/ML brain-computer interface startup running stateful transformations for AI agent pipelines at 7 TB/day with a 93% reduction in silently duplicated or corrupted metrics) and Restack (AI agent platform observability at ~10 TB/day and ~6.7B messages/day, citing an 84% reduction in engineering effort and an 87% reduction in pipeline-related incidents).

The company has raised approximately $5.9M cumulatively — a $1.1M pre-seed in November 2023 led by High-Tech Gründerfonds with Roosh Ventures, and a €4.43M (~$4.8M) seed in October 2024 led by Upfront Ventures with HTGF, Robin Capital, TinyVC, and angels from Weaviate, Redis, Sentry, and Snowflake. Headcount is reported in the 1–10 range, with named team members spanning engineering (senior frontend, backend, solution), operations, marketing, and project management. Product release cadence has accelerated through 2025–2026, with v2.0.0 (Sept 2025, 282 commits), v2.2.0 (OpenTelemetry, Oct 2025), v2.4.0 (Kafka Kerberos, Nov 2025), and v3.x releases in early 2026, alongside demonstrated horizontal scaling to 500k+ events/sec. The company is registered as a GmbH in Berlin under HRB252594 B and operates under a clear mission to "democratize data streaming" for the next generation of data teams.

Name

GlassFlow

Legal name

GlassFlow GmbH

Website

https://glassflow.dev

Company type

Private

Founded year

2023

Operating status

Operating

Headcount range

1–10 employees

Short description

GlassFlow is a Berlin-based open-source streaming ETL platform that ingests data into ClickHouse with built-in CDC, deduplication, and schema evolution, targeting Python data engineers and AI/ML observability workloads.

Ownership category

akta.pro rank

Product category

Data Streaming Infrastructure

NAICS

Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (518210)

SIC

Services-Prepackaged Software (7372)

akta.pro primary industry

Real-Time / Streaming Data Warehousing (HDAEABAF)

Keywords

• Data streaming platform
• ClickHouse ingestion
• Streaming ETL
• CDC pipelines
• Event-driven pipelines

Headquarters

HQ city

Berlin

HQ country

Germany

HQ region

Europe

Offices1 record

Markets served

GTM type

B2B

Offering type

Software

Cost components

Technology or R&D, Personnel, Infrastructure, Operations, Marketing or Sales

Revenue model

1. Managed Cloud Service (BYOC): GlassFlow runs in the customer's cloud account with GlassFlow managing upgrades, infrastructure, reliability and operations. Data stays within customer's infrastructure. This is the primary managed service offering with production reliability and support.
2. Self-Hosted Deployment: GlassFlow runs entirely in customer infrastructure with full control over networking, security, and scaling. Customers manage their own deployment in Kubernetes.
3. Enterprise Support & SLAs: 24/7/365 support, SLAs, Slack/Teams support, personalized onboarding and training, and implementation engineers available as add-on services for enterprise customers.

Pricing tiers

Go-to-market motion2 records

Distribution channels4 records

Marketing channels7 records

Core offering

GlassFlow is a serverless, event-driven data streaming platform purpose-built for ClickHouse ingestion. It enables Python developers and data engineers to build, manage, and scale event-driven data pipelines with built-in CDC for PostgreSQL/MySQL/MongoDB, schema evolution handling, deduplication, data normalization, and high-throughput ingestion into ClickHouse. The platform is offered as open-source software, a managed cloud service (BYOC), and a self-hosted enterprise deployment with Kubernetes-native architecture.

Product overview

GlassFlow is an open-source streaming ETL platform designed specifically for ClickHouse ingestion. The product portfolio consists of the core GlassFlow platform plus multiple deployment options and add-on modules. The core platform provides serverless execution, pre-built connectors, and Python-native SDK for building event-driven data pipelines. Deployment options include a Managed Cloud Service (BYOC) where GlassFlow runs in the customer's cloud, and Self-Hosted Kubernetes deployment for full infrastructure control. Key modules include Managed Kafka Connector for enterprise Kafka clusters, CDC capabilities for PostgreSQL/MySQL/MongoDB, Deduplication Engine with configurable time windows, Data Transformation for normalization and joins, Dead-Letter Queue for fault handling, and OpenTelemetry integration for observability data. The platform enables companies to create pipelines in approximately 15 minutes and scales to 500k+ events per second.

Differentiator

Problem solved

Functional benefit

Products and services

• GlassFlow Streaming ETL Platform Open-source streaming ETL platform purpose-built for ClickHouse ingestion. Provides serverless execution, pre-built source connectors (Kafka, PostgreSQL, MySQL, MongoDB, OpenTelemetry, AWS services), Python SDK, CDC ingestion, deduplication, schema evolution handling, data normalization, and horizontal scaling to 500k+ events/sec. Targets Python data engineers and companies running ClickHouse for real-time analytics, observability, and AI/ML workloads.
• clickhouse-etl (Open Source) Open-source repository on GitHub containing the ClickHouse ETL implementation, enabling developers to self-host the full GlassFlow core functionality at no cost for free tier usage.
• Managed Cloud Service (BYOC) Fully managed GlassFlow service where the platform runs in the customer's own cloud account with GlassFlow managing upgrades, infrastructure, reliability, and operations. Data remains within the customer's infrastructure while GlassFlow handles the operational overhead.
• Self-Hosted Deployment Kubernetes-native self-hosted deployment option for teams requiring full control over networking, security, and scaling. Customers manage their own deployment in Kubernetes with optional enterprise support and SLAs as add-ons.
• Enterprise Support & SLAs Premium enterprise support add-on providing 24/7/365 support, SLAs, Slack/Teams support channels, personalized onboarding and training, and dedicated implementation engineers for enterprise customers.

Quantifiable outcome

• 84% reduction in engineering effort for building and maintaining pipelines
• +5 more outcomes

Named customers2 records

Segments3 records

Ideal customer profiles2 records

Technology focussed Yes

API detail

Has API

No

API docs

API detail

Core technology

AI maturity

App detail

Integration32 records

Feature10 records

Scale indicators7 records

Recent moves6 records

Expansion highlights6 records

Direct peers

• Decodable: Was a direct streaming ETL/CDC competitor offering a managed platform for moving data into analytical databases. Acquired by ClickHouse in 2024, meaning it now competes with GlassFlow as part of ClickHouse's own first-party ingestion story.
• Estuary Flow: Open-source and managed streaming ETL platform supporting real-time CDC and materializations into analytical stores. Closely comparable to GlassFlow in product shape (open-source core, managed cloud, connectors, CDC) and target customer (data engineering teams).
• Materialize: Streaming SQL engine that performs real-time transformations on data in motion, targeting the same Python/data-team workloads that struggle with Flink. Directly comparable as a real-time data transformation and materialization platform.
• RisingWave: Streaming database with SQL-based stream processing and materialization, designed to be simpler than Flink. Targets the same developer-friendly streaming analytics use case as GlassFlow.
• Airbyte: Open-source data integration platform with a large connector catalog and self-hosted/managed deployment options. Overlaps with GlassFlow in open-source ETL distribution, though broader in destination support.
• Debezium: Open-source CDC platform that captures row-level changes from databases. Directly comparable in CDC capability — GlassFlow actually supports Debezium-format streams — and competes for the same change-data-capture workload.

Emerging players

• Conduit: Open-source streaming data integration platform focused on real-time CDC and connectors. Comparable in architecture and target user to GlassFlow, though at an earlier stage of maturity and feature set.

Broad incumbents

• Confluent: Commercial vendor behind Apache Kafka and ksqlDB; the dominant general-purpose streaming platform. Overlaps with GlassFlow's Kafka ingestion capabilities but serves a much broader, Kafka-centric customer base.
• Apache Flink: General-purpose distributed stream processing framework. The incumbent alternative GlassFlow explicitly positions against for Python teams unwilling to operate Java/Scala-heavy infrastructure.
• Striim: Enterprise real-time data integration platform offering CDC, streaming ETL, and analytics pipeline capabilities. Comparable in CDC and streaming ingestion features, but at a larger enterprise scale than GlassFlow.

Market position

Strengths4 records

Weaknesses4 records

Competitive moat4 records

Key risks5 records

Key highlights7 records

Customer concentration

• Xx.com/glassflowdev
• GitHubgithub.com/glassflow/clickhouse-etl
• LinkedInlinkedin.com/company/glassflow-dev

Revenue estimate

Valuation estimate

Number of profiles

Profiles9 records

Funding overview

Funding rounds2 records

Investors7 records

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about GlassFlow