Description
Chronos DB is a specialized database built for the autonomous AI era, offering a robust solution for managing data interactions with AI agents. It is fully compatible with Postgres, meaning existing applications, ORMs like Prisma and Drizzle, and tools such as psql and psycopg can connect and operate without modification. The core innovation of Chronos lies in its ability to provide every AI agent with its own isolated "world" – an instant copy of the entire production database. This allows agents to experiment and make changes without risking the integrity of the main production data.
Read More
The system enforces strict rules for merging changes back into the main database. Users can define merge policies that dictate which changes an agent may land on its own (e.g., "at most 500 rows, no deletes, never the payments table"). Changes that comply with these rules merge automatically, while those that break any rule are held for human review and approval, complete with clear reasons for the hold. This mechanism prevents accidental data corruption, a common challenge when AI agents interact with live production environments.
Chronos also incorporates a powerful history feature, allowing users to read any table as it was at any minute within the last hour. Furthermore, any merged change, or even all actions performed by a specific agent, can be undone row by row for an hour after it lands. This provides an essential safety net, enabling quick rollbacks in case of unintended consequences. The database is designed for high performance, particularly for agent-centric workloads, demonstrating significantly faster performance than standard Postgres or Dolt in scenarios involving numerous agents copying, writing, and merging data.
Chronos DB is source-available under the BUSL-1.1 license and can be self-hosted for free. A hosted Chronos Cloud service and a Hosted Gateway (for existing Postgres installations) are planned for future release, offering managed solutions. It integrates seamlessly with popular development tools and languages such as LangChain, Python, Node.js, SQLAlchemy, Supabase, and Neon, and includes built-in pgvector types for vector embeddings. The system is developed by 2090 Labs, a team focused on building foundational technologies for the future of AI.
- CategoryAI SQL & Database Tools
- Pricing ModelFree (Self-Hosted)
- DeploymentSelf-Hosted
- PlatformsWeb
Chronos DB's Core Features
Instant Data Copies for Agents (Worlds)
Postgres-Compatible Wire Protocol
Configurable Merge Rules and Policies
Human Approval for Risky Changes
Time Travel for Data (Read Past States)
Undo Merges Row by Row (within 1 hour)
Built-in Vector Types (pgvector)
MCP Server for Agent Interactions
Self-Hosted Option (Free)
Source-Available (BUSL-1.1)
Integration with ORMs (Prisma, Drizzle)
Support for Python and Node.js
Benchmarked for Agent Workloads
Signed Releases for Security
How to use Chronos DB?
Install Chronos DB: Run a single line command to install Chronos DB on your machine.
Connect Agents: Configure your AI agents to connect to Chronos DB using the Postgres wire protocol or the MCP server.
Define Merge Rules: Set up merge policies to specify which changes agents can land automatically and which require approval.
Develop with Worlds: Allow agents to work in their own isolated copies ("worlds") of your production data.
Review and Approve: Monitor agent changes in the dashboard and approve or reject merges that break predefined rules.
Utilize History: Read past states of any table or undo recent merges if needed.
Chronos DB's Use Cases
- Safe AI Agent Development
- Automated Data Updates
- Human-in-the-Loop Approvals
- Data Rollbacks and Recovery
- AI Agent Training
- Experimentation with Data
- Complex Data Migrations
- Real-time Data Investigation
