Description
ObsessionDB offers fully managed hosting for ClickHouse software, providing a high-performance and scalable data warehouse solution optimized for AI workloads. It functions as a drop-in replacement for existing ClickHouse deployments, ensuring compatibility with current queries, integrations, and tooling without modification. The platform is designed to deliver superior performance and cost efficiency, particularly as data volumes grow, outperforming many providers at 100 TB compared to their 1 TB capabilities.
Read More
Users have the flexibility to deploy ObsessionDB in their own AWS or Google Cloud account (BYOC) or on their own hardware, with the ObsessionDB team managing operations in every scenario. This BYOC model ensures data remains within the user's cloud environment, allowing them to retain existing infrastructure discounts while ObsessionDB charges an annual fee for license, operations, and support. Alternatively, ObsessionDB can provide fully managed hosting across AWS, Google Cloud, and Azure in every public region.
At its core, ObsessionDB utilizes the official ClickHouse engine but rebuilds the storage layer for compute-storage separation. This proprietary architecture, featuring the SharedMergeTree engine, a distributed NVMe cache ring (Bunsen), and locality-aware merge and read routing (Affinity), allows data to reside on object storage while achieving local-disk speeds. This design ensures that compute nodes are stateless and interchangeable, enabling scaling of reads by adding compute without re-sharding or rebalancing data.
The platform is particularly well-suited for agentic workloads, which involve relentless, parallel probing across data. ObsessionDB provides isolated, read-only compute for agents, preventing them from impacting production traffic. It ensures consistent query speeds even after hundreds of probes, with every agent action logged for accountability. The pricing model for agents is based on compute hours, not per query, encouraging extensive exploration. Additionally, CHKit offers agents structured commands to operate the database, including schema and pipeline deployment as code, and allows for sandboxed experimentation with production data branches.
ObsessionDB is enterprise-ready, offering dedicated migration teams, enterprise SLAs with uptime commitments, and robust failure recovery mechanisms. Data durability is ensured by storing row data in S3, with point-in-time recovery capabilities. Security features include single sign-on (SSO), granular access control down to table and column levels, private networking, dedicated compute, and encryption in transit and at rest. A 24/7 on-call team provides expert support for critical issues.
- CategoryAI SQL & Database Tools
- Pricing ModelUsage-based
- DeploymentCloud, BYOC, On-Premise
- PlatformsWeb
ObsessionDB's Core Features
Fully managed ClickHouse hosting
BYOC (Bring Your Own Cloud) deployment on AWS, Google Cloud, Azure, or own hardware
Compute-storage separation with ClickHouse engine
Optimized for agentic AI workloads
Distributed NVMe cache ring (Bunsen)
Locality-aware merge and read routing (Affinity)
Isolated, read-only compute for AI agents
No per-query charges for agent workloads
CHKit for agent database operations and schema deployment
Sandbox environments per agent for experimentation
Dedicated migration engineering team
Enterprise SLAs and uptime commitments
Point-in-time recovery
Single sign-on (SSO) and granular access control
Private networking and dedicated compute
24/7 on-call support
How to use ObsessionDB?
Explore Features: Review the ObsessionDB website to understand its capabilities, architecture, and how it differs from standard ClickHouse deployments.
Calculate Costs: Use the online pricing calculator to estimate monthly costs based on your storage, memory, and node requirements, comparing it against other managed ClickHouse solutions.
Start Free Tier: Spin up a free developer instance to test your workloads and kick the tires with a low-to-medium sized cluster.
Contact Sales: For BYOC deployments, custom quotes, or heavier workloads, contact the sales team to discuss a proof of concept.
Migrate Data: Engage the dedicated migration team to plan and execute the migration of your existing ClickHouse workloads onto ObsessionDB.
Connect Tools: Integrate ObsessionDB with your existing language drivers, BI tools, streaming ingestion platforms, and AI agents using provided guides and standard ClickHouse protocols.
Monitor Performance: Utilize the built-in monitoring dashboard to track query performance, errors, storage, and resources, or query the raw observability surface directly.
Scale Resources: Resize your cluster at any time by adding or removing nodes, changing compute size, or adjusting storage to match evolving workload demands.
ObsessionDB's Use Cases
- High-Performance Data Warehousing
- AI Agent Workloads
- Real-time Analytics
- Cost-Optimized ClickHouse Deployment
- BYOC Data Management
- Data Ingestion and ETL
- Business Intelligence Dashboards
- Log and Metrics Analysis
- Schema and Pipeline Deployment
- Secure Data Environments







