Feature Requests

Added a visual diagram of a deployment's schema on the Convex dashboard. View it on the Schema page: https://dashboard.convex.dev/deployment/schema

This feature will allow configuring granular usage limits for individual deployments. e.g. "Allow 50 GB of DB I/O per month"

When running a query or mutation with runQuery/runMutation, the subtransaction can currently do as many reads/writes as it wants. It would be useful to be able to set a fixed limit on how much reads/writes the subtransaction can do.

The Convex cloud now avoids invalidating all queries on code push. Instead, it relies upon a more granular analysis of what functions actually changed to make it so that most cached queries stay warm in the cache and only truly invalidated things are recalculated.

Change the default behavior of preview deployments to re-use rather than creating a new deployment on each build. Allow re-creating by explicitly passing --preview-create

Allow passing a --message to npx convex deploy, and provide defaults from common CI environments

Allow setting and getting default environment variables via npx convex env default get/list/set

Surface the currently executing function's name, component, function type (query/mutation/action) and visibility.

Allow queries and mutations to inspect how much bandwidth has been consumed so far, so they can stop early or observe how much bandwidth sub-transactions took

Support snapshot queries that provide a stale, cached query value and do not take any dependencies on read records. These enable high-performance "queue tailing" and related features for implementing efficient components like Workpool.

Rather than requiring an explict re-export from the app, allow the app to provide a mount root that components can use to create their own HTTP endpoints.

It would be helpful if the command line allowed developers to better manage multiple local and cloud development environments.

Give the command line the power to create time-limited deployments, similar to preview deployments, for any period developers want.

!cache chart We shipped a new dashboard feature that helps you understand how cache invalidations work in Convex. If you've ever wanted to optimize or understand your cache usage, this is for you!

Integrate PostHog's LLM analytics capabilities with the convex agent component. Docs Example

Can you please update the dashboard usage to show the current usage vs the total limit please? For example: I would like to see the current usage vs the total usage available, for example on free plan: Function calls: 1.3K/1M This helps to keep track of usage and be aware of getting closer to the cu...

Support industry-standard audit logging for compliance needs. Will work even with cached queries. Recorded into an S3-compliant bucket for integration with the customer's SIEM tooling.

New npx convex deployment create and npx convex deployment select commands for creating cloud deployments and selecting which deployment to target. All CLI commands now support --deployment flag accepting deployment names, refs (e.g. dev/james), dev, prod, or cross-project refs (team-slug:project-sl...

The CLI can now automatically provide Convex AI context files (AGENTS.md, CLAUDE.md, guidelines/state files) to your project when using npx convex dev. Manage them with npx convex ai-files: install, update, status, disable, enable, and remove.

npx convex init now supports non-interactive agent workflows via CONVEXAGENTMODE=anonymous. Combined with npx convex env set, agents can fully initialize deployments and configure environment variables without human interaction.