Anyquery
by julien040
Query any data source with SQL — then expose it via MCP
database Go Intermediate Self-hostable No API key
⭐ 1.6k stars 📅 Updated: 3w ago
Description
Anyquery is a SQL query engine that can talk to virtually anything: Airtable, Google Sheets, Notion, GitHub, CSV files, JSON APIs, and dozens more. Its MCP integration means your AI agent can run SQL queries against these sources conversationally, without you having to set up a database or write integration code. Think of it as a universal data adapter that speaks SQL and MCP.
The clever design choice is treating every data source as a virtual SQLite table. Your agent can JOIN a Google Sheet with a GitHub issue list, or filter Airtable records using standard SQL WHERE clauses. This is genuinely powerful for ad-hoc data exploration where the data lives across multiple SaaS tools and you do not want to build a pipeline just to answer a question.
Written in Go, Anyquery is a single binary with no dependencies, which makes setup painless. The plugin system supports 40+ data sources, and adding new ones follows a documented pattern. The main limitation is that write operations are limited to some plugins — most sources are read-only through Anyquery.
✅ Best for
Developers and data analysts who need to query data scattered across multiple SaaS tools
⏭️ Skip if
You need a production data pipeline with writes — use a proper ETL/database setup instead
💡 Use cases
- Query Notion databases, Google Sheets, and Airtable using SQL from your AI agent
- Join data across SaaS tools without building ETL pipelines — "combine GitHub issues with Jira tickets"
- Explore CSV/JSON datasets interactively by asking your agent SQL-based questions
👍 Pros
- ✓ Supports 40+ data sources via plugins, all queryable with standard SQL
- ✓ Single Go binary, no runtime dependencies — installs in seconds
- ✓ Virtual table approach lets you JOIN across completely different data sources
👎 Cons
- ✗ Most plugins are read-only — limited write-back capability to source systems
- ✗ Complex queries across multiple remote sources can be slow due to API pagination
- ✗ Plugin quality varies — some community plugins are less maintained than core ones
💡 Tips & tricks
Start with the CSV and JSON plugins to test locally before connecting SaaS sources.
Use `anyquery -q "SELECT * FROM ..."` to verify queries work before exposing them
via MCP. For Google Sheets, the setup requires OAuth — do this once interactively
and Anyquery saves the token for headless use.
Quick info
- Author
- julien040
- License
- NOASSERTION
- Runtime
- Go
- Transport
- stdio
- Category
- database
- Difficulty
- Intermediate
- Self-hostable
- ✅
- API key
- No API key needed
- Docker
- —
- Version
- 0.0.0
- Updated
- Jan 27, 2026
Client compatibility
- ❓ Claude Code
- ❓ Cursor
- ❓ VS Code Copilot
- ❓ Gemini CLI
- ❓ Windsurf
- ❓ Cline
- ❓ JetBrains AI
- ❓ Warp
Platforms
🍎 macOS 🐧 Linux 🪟 Windows