Duckle 0.2.0 speeds up local data workflows
Duckle, an open-source local data tool, has a new 0.2.0 release. It adds faster ways to move, join, and update large datasets on a normal laptop. This can help teams prepare fresher data for AI agents without rerunning everything.
Duckle lets people build data pipelines on a visual canvas and run them locally with DuckDB. The new version adds CDC, incremental loads, parallel pipelines, and a Visual Map feature for joining data without writing SQL. The post says it works across CSV, Parquet, DuckDB, SQLite, warehouses, relational databases, and MongoDB.
The post claims the tests ran on a plain 16 GB laptop. A 16-node pipeline with 5 million rows, a 3-way join, parallel branches, and 4 outputs took about 3.0 seconds. A 100,000-row DuckLake CDC mirror took about 1.7 seconds, and a 5 million-row incremental load took about 1.8 seconds. For AI agent builders, the practical value is data prep: update only what changed before sending cleaner, smaller context to models.
Key points
- Duckle is a free, open-source, local-first data studio.
- Version 0.2.0 adds CDC, incremental loads, parallel pipelines, and Visual Map.
- The post reports 5 million-row jobs finishing in roughly 2 to 3 seconds on a 16 GB laptop.
- It says there is no cloud account, no telemetry, and data stays on the user’s machine.
- For AI agents, this may lower prep time and reduce unnecessary data sent into models.
Quick term guide
- open-source
- Software whose code is shared publicly so others can inspect, use, or change it.
- AI agents
- AI agents are AI tools that can carry out steps toward a goal, not just answer once.
- AI agent
- An AI program that can inspect information and suggest what to do next.
- data pipelines
- A set of steps that move, clean, combine, or export data.
- data pipeline
- An automated path that moves data from where it's collected to where it's needed, such as an AI model.
- incremental loads
- Updating only the changed data instead of loading the whole dataset again.
- database
- A large collection of organized data used for search and analysis.
- local-first
- An app design where your data is mainly stored and controlled on your own device.