Hi @everyone! Time for an update on our roadmap for the next 6 weeks. 🧑‍🤝‍🧑 **Partnerships** Following the Chef launch there have been a lot of companies reaching out interested in building apps on Convex. We want to make this as easy as possible. - Seamless API for partners to deploy apps on behalf of customers. This is already possible but undocumented so reach out if you’re interested! - Account and billing changes to allow Convex to show up in third-party marketplaces and codegen tools. - Log in with a non-github identity. :chef: **Chef** Keep leveling up what is possible in codegen via the power of good APIs. - Convex Components support. Soon you’ll be able to say something like “build a collaborative text editor with presence” and it’ll work with ProseMirror, etc - Token drops and better token utilization. - Cheaper plan for Chef-only users. 📈 **Scale** Our usage numbers have been going up a lot. We’re scaling fine with project count but we’d like to improve performance for individual deployments. - Increase module size limits. - Increase transaction limits, documents read/written etc. - Decrease function latency, especially latency for calling components. - Faster typescript analyze time in the dev loop. 🧑‍🎓 **Graduate existing features** Some of our features have been in beta for too long. - Ship officially-supported Better Auth integration, take some version of Convex Auth out of beta. - Fivetran integration out of beta. - Our Convex Components out of beta. - Simplify Convex Helpers for common needs like joins, pagination over multiple queries, etc, surface them to be easier to use. 🤖 **AI/MCP** Make the backend for AI better at building backend for/with AI. - The Agent component has been getting popular so we have some more tweaks to do to that. - Release profiling MCP endpoint that will let a model figure out if a query is efficient or not. - Support for calling MCP over http etc instead of just from cli. ♻️ **Sync** “Local first”/local store has been on hold since we haven’t been happy with the API. We’re revamping the project. - Client side ID allocation. - More efficient readset intersection (faster subscriptions). - Revised local sync client that’s a lot easier to use than our old prototype with resolvers. 🤷 **What’s not on here?** Lots of stuff! We already have built-in web hosting for Chef apps and it wouldn't be much of a stretch for us to add that as a first class feature to Convex. We're considering moving some backend storage to Planetscale Metal, will be interesting to see fi there's a performance win there. There have also been a lot of requests for larger-scale Convex self-hosting or running it in a customer VPC so would love to have a great solution there.

james•6mo ago

Hi @everyone! Time for an update on our roadmap for the next 6 weeks. 🧑‍🤝‍🧑 Partnerships Following the Chef launch there have been a lot of companies reaching out interested in building apps on Convex. We want to make this as easy as possible. - Seamless API for partners to deploy apps on behalf of customers. This is already possible but undocumented so reach out if you’re interested! - Account and billing changes to allow Convex to show up in third-party marketplaces and codegen tools. - Log in with a non-github identity. :chef: Chef Keep leveling up what is possible in codegen via the power of good APIs. - Convex Components support. Soon you’ll be able to say something like “build a collaborative text editor with presence” and it’ll work with ProseMirror, etc - Token drops and better token utilization. - Cheaper plan for Chef-only users. 📈 Scale Our usage numbers have been going up a lot. We’re scaling fine with project count but we’d like to improve performance for individual deployments. - Increase module size limits. - Increase transaction limits, documents read/written etc. - Decrease function latency, especially latency for calling components. - Faster typescript analyze time in the dev loop. 🧑‍🎓 Graduate existing features Some of our features have been in beta for too long. - Ship officially-supported Better Auth integration, take some version of Convex Auth out of beta. - Fivetran integration out of beta. - Our Convex Components out of beta. - Simplify Convex Helpers for common needs like joins, pagination over multiple queries, etc, surface them to be easier to use. 🤖 AI/MCP Make the backend for AI better at building backend for/with AI. - The Agent component has been getting popular so we have some more tweaks to do to that. - Release profiling MCP endpoint that will let a model figure out if a query is efficient or not. - Support for calling MCP over http etc instead of just from cli. ♻️ Sync “Local first”/local store has been on hold since we haven’t been happy with the API. We’re revamping the project. - Client side ID allocation. - More efficient readset intersection (faster subscriptions). - Revised local sync client that’s a lot easier to use than our old prototype with resolvers. 🤷 What’s not on here? Lots of stuff! We already have built-in web hosting for Chef apps and it wouldn't be much of a stretch for us to add that as a first class feature to Convex. We're considering moving some backend storage to Planetscale Metal, will be interesting to see fi there's a performance win there. There have also been a lot of requests for larger-scale Convex self-hosting or running it in a customer VPC so would love to have a great solution there.

0 Replies

No replies yetBe the first to reply to this messageJoin

Did you find this page helpful?