Anthropic recognized PostgreSQL backup tool Databasus as a valuable open source project

#ai #postgres #opensource #database

Anthropic, the company behind Claude AI, accepted Databasus into their Claude for Open Source program. Databasus is the most widely used PostgreSQL backup tool on GitHub and an industry standard for PostgreSQL backup management. The program gives maintainers of significant open source projects free access to the highest-tier Claude subscription. Here's what happened, how the program works and why it matters for the open source ecosystem.

How Databasus got here

Databasus started in 2023 as a small internal tool called Postgresus. The original author built it to manage PostgreSQL backups across multiple production and personal projects. The problem it solved was familiar to anyone running databases: writing bash scripts around pg_dump, setting up cron jobs, configuring storage and wiring up notification webhooks for each database separately.

The tool went open source in early 2025. Within months it gained significant traction. By the end of 2025 it had become the most starred PostgreSQL backup tool on GitHub, surpassing established projects like WAL-G, pgBackRest and Barman. The project was renamed to Databasus in late 2025 after expanding to support MySQL, MariaDB and MongoDB.

As of early 2026, the project has around 5,800 GitHub stars, over 250,000 Docker pulls and tens of thousands of daily active users. The user base includes individual developers, DevOps teams, DBAs and enterprises running it in production.

What Databasus does

Databasus is a free, open source, self-hosted backup management tool for PostgreSQL, MySQL, MariaDB and MongoDB. It runs as a single Docker container under the Apache 2.0 license. All data stays on the user's infrastructure.

Feature	Details
Scheduling	Hourly, daily, weekly, monthly or custom cron
Storage destinations	Local disk, S3, Cloudflare R2, Google Drive, Dropbox, SFTP, NAS and 70+ providers via Rclone
Encryption	AES-256-GCM with unique keys per backup file
Retention policies	Time period, count, size limit or GFS (Grandfather-Father-Son) for compliance
Health monitoring	Configurable failure thresholds with alerts
Team features	Workspaces, RBAC (viewer/member/admin/owner), audit logs
Notifications	Slack, Discord, Telegram, MS Teams, Email and webhooks
Cloud databases	AWS RDS, Google Cloud SQL, Azure — something pgBackRest and Barman can't do
Backup portability	Decrypt and restore with only secret.key and standard tools, no Databasus required

The tool is suitable for both individual developers backing up side projects and enterprise teams managing dozens of production databases with strict compliance requirements.

What is the Claude for Open Source program

Anthropic's Claude for Open Source program provides eligible open source maintainers with 6 months of free Claude Max 20x — the highest-tier consumer Claude subscription with 20x the usage limits of Claude Pro. The program accepts up to 10,000 recipients total, with applications closing June 30, 2026.

There are two eligibility tracks with different requirements.

Track	Requirements
Maintainer Track	Primary maintainer of a public GitHub repo with 5,000+ stars or 1M+ monthly NPM downloads. Must show active maintenance within the last 3 months and have merge access
Ecosystem Impact Track	For projects below the star threshold but with meaningful ecosystem dependence. Anthropic evaluates downstream dependents, breadth of usage and criticality

Applications are reviewed on a rolling basis. Meeting the thresholds doesn't guarantee acceptance. Anthropic approves or denies at their discretion. The requirements are not low — they filter for projects that people actually depend on in production.

Why this recognition matters

Databasus qualified through the Maintainer Track. The star count exceeds the 5,000 threshold, the project is under active development and the adoption numbers are substantial.

For an open source project that handles something as sensitive as database backups, external validation from a company like Anthropic carries weight. When teams evaluate whether to trust a tool with their production data, signals like program acceptance from a reputable organization help build confidence. It's not just about the subscription itself — it's about what the selection process represents.

This is also a positive signal for the broader open source community. Programs like this show that companies are willing to invest in the tools that developers and teams rely on daily.

The project's approach to AI in development

One interesting aspect of Databasus is its transparent stance on AI usage. The project has a dedicated AI disclaimer in the README that explains exactly how AI is and isn't used in development.

The maintainer uses Claude as a development tool for code quality verification, vulnerability scanning, documentation cleanup and double-checking PRs after human review. But the project explicitly prohibits using AI to write entire code, "vibe coding" and shipping code without line-by-line human verification and tests.

The project has solid test coverage with both unit and integration tests, a CI/CD pipeline with automated checks and manual PR verification. Vibe coded contributions are rejected by default. There have been documented cases where PRs were denied specifically because they were obviously AI-generated without proper review.

This responsible approach to AI usage aligns well with what a credible backup tool should demonstrate. People trust the project with their production data, and that kind of trust requires engineering rigor rather than shortcuts.