Feature Requests

Allow using the AI Gateway through EU servers.

This feature will allow configuring Custom Roles to only have access to specific components, instead of to all data in the entire application

Certain workloads like generating a PDF or processing/zipping larger files are currently impossible with the low memory limits on actions. It would be ideal to at least have the option to specify the memory limit on actions (ideally per action like AWS Lambda). Even 1GB would be better than nothing,...

It's very useful to be able to view traces and spans for operations in logging applications like Axiom. Currently, there are some limitations that make this difficult to implement, and in some cases, impossible: The OpenTelemetry node packages require "use node", so the only place we can use the off...

Convex currently enforces its own _id primary key for all tables which is problematic and sub-optimal performance-wise in various use cases: Migrating data from previous databases to preserve primary keys/relationships Using 3rd parties as the source of truth for primary keys e.g. Clerk users, orgs,...

Provide caching at the "sub query" level inside the convex deployment. Specifically if a developer calls ctx.runQuery(...) within a query or mutation, and that individual query's cache value is still valid, the cached value will be used instead of running v8 again. This will allow developers to cont...

To better handle background jobs

Convex cloud provides global edge caches that serve cached queries at low latency everywhere on earth.

Support inexpensive copy-on-write database clones so teams can create realistic test environments and reproduce bugs without paying full storage costs each time.

Deliver the next iteration of Chef with a more capable developer workflow, clearer abstractions, and stronger support for repeatable setup and orchestration.