Rerun is building a visualization engine for streams of multimodal data that's easy to use, adapt, and extend.
This page is meant to give an high level overview of ongoing and planned work.
- Performance improvements
- UX & DX improvements
- Supporting more data types
- Improvements to the C++ SDK
- Hosted api reference documentation
- Even easier cmake build system integration
- Logging non-rerun types is faster and easier
- Expand the "Visual History" feature to "Visible Time Range"
- For supported views and data types you'll be able to specify both absolute and relative time ranges to include data from.
- Significant performance improvements for all time range visualizations, particularly time series.
- End to end performance for high frequency time series logging
- Layout and configuration from code (blueprint)
- Datasets that are bigger than RAM for the native viewer
- Extensibility of all parts of the stack
- Easily read query data recordings from user code
- Callbacks and the ability to build interactive applications with Rerun
- For example: UI for tweaking configs, custom data annotation tools, etc
- Data format stability