# Loony

> Loony is a CLI-first managed backend for pulling data from APIs, keeping it synced on a schedule, and exposing it via REST or MCP. Describe your data sources, deploy in 30 seconds, query from anywhere. No infrastructure to manage, no forced schemas.

## What Loony does

Loony replaces the duct-tape stack of Postgres + cron jobs + Flask APIs that developers cobble together to keep API data fresh and queryable. One CLI command provisions a managed Postgres database, runs your data pipelines on a schedule, and exposes the results as REST and MCP endpoints.

### How it works

1. **Describe** -- Tell your agent (or write yourself) what data you need. Any API source, any schema. Your agent writes the scripts, transforms, and schema.
2. **Deploy** -- Run `loony deploy`. Loony validates the schema, provisions the database, runs scripts, registers schedules. Takes ~30 seconds.
3. **Query** -- Your data is available via REST API, MCP endpoint, and direct Postgres connection. Agents query via MCP tools. BI dashboards connect directly.

### Key capabilities

- **Agent-native**: Every project ships with `.claude/rules`. Describe what data you need; your agent builds, tests, and deploys the pipeline.
- **Managed Postgres**: Every app gets its own database. Zero config. Data accumulates over time.
- **REST + MCP endpoints**: Authenticated endpoints with full OpenAPI spec. Query from anywhere. MCP endpoints let AI agents access your data as tools.
- **Scheduled pipelines**: Cron per script. Runs on Loony infra. Pause, resume, or change anytime. Schedules from daily (free) to real-time (Business).
- **Built-in semantic layer**: Opinionated schemas with field-level comments optimised for zero hallucinations when queried by LLMs.
- **BYODB**: Bring your own database on Pro and Business plans.

### Who this is for

- AI agent builders who need structured, fresh data queryable via MCP
- Indie hackers and solo developers building data products (dashboards, alerts, APIs) without managing infra
- Small data teams at startups consolidating API data without enterprise ETL tools

### What Loony is NOT

Loony is not an enterprise ETL tool (like Fivetran or Airbyte with 400+ connectors). It is not an app backend (like Supabase). It is not a workflow automation platform (like Pipedream or Zapier). It is a personal/team data backend: pull, store, transform, serve -- the full loop in one tool.

## Install

```
pip install loony
```

## Pricing

- **Free**: $0/mo -- 1 app, Every minute schedules, 200 runs/mo, 500MB managed DB, 7-day log retention, Community support
- **Hobby**: $12/mo -- 3 apps, Every minute schedules, Unlimited runs, 5GB managed DB, 30-day log retention, Email support
- **Pro**: $59/mo -- 5 apps, Every minute schedules, Unlimited runs, 20GB managed DB, BYODB (any database), 90-day log retention, Email support
- **Business**: $179/mo -- 20 apps, Real-time schedules, Unlimited runs, 100GB managed DB, BYODB (any database), 1 year log retention, Dedicated support
- Extra apps on paid plans: $10/mo each

## Pages

- [Homepage](https://loony.dev/): Product overview, how it works, features, live examples, pricing
- [Data Processing](https://loony.dev/data-processing): Data processing agreement
- [Integrations](https://loony.dev/integrations): Supported data source integrations
- [Privacy](https://loony.dev/privacy): Privacy policy
- [Terms](https://loony.dev/terms): Terms of service

## Optional

- [Contact](mailto:hello@loony.dev): For custom SLAs, database integrations, or enterprise plans