Make Convex more composable/extensible by providing support for self-contained components that run in their own namespace but within the same transactional context as the caller.
Rearchitect our internal systems for executing user functions to significantly increase function parallelism.
Open source the core of the Convex backend so it can be run locally by developers as a single binary.
Support backup and restore for Convex projects while maintaining referential integrity for document IDs.
Start enforcing limits on free accounts and charging customers for usage on pro accounts, with generous additional allowances.
Scale up vector search to be well ahead of customer workloads. Add prefix search and fuzzy search to full-text search.
This wasn’t mentioned in our last post but the plan was to get Convex ready for our SOC-2 audit.
Overall I’d give us a medium rating. We took on a number of fairly open ended projects with longer timelines and made good progress but going into the next quarter we want to tighten up execution a little and prioritize improving the experience for existing Convex customers with growing applications.
