Which auth to pick
I'm about to start a new (big) project with Convex and Tanstack Start. What auth should I pick? Convex Auth or Better Auth (or any other)?
10 Replies
Thanks for posting in <#1088161997662724167>.
Reminder: If you have a Convex Pro account, use the Convex Dashboard to file support tickets.
- Provide context: What are you trying to achieve, what is the end-user interaction, what are you seeing? (full error message, command output, etc.)
- Use search.convex.dev to search Docs, Stack, and Discord all at once.
- Additionally, you can post your questions in the Convex Community's <#1228095053885476985> channel to receive a response from AI.
- Avoid tagging staff unless specifically instructed.
Thank you!
Depends what your auth requirements are. If convex auth covers your needs, it'll like be faster to integrate than anything else. You can always change down the line.
Correct me if I'm wrong, but maybe I have read somewhere that convex auth will be deprecated in favour of better auth (or at least Convex will recommend Better Auth as primary option)?
I think it's an open question bc the underlying lib (authjs?) was acquired by better auth. Better Auth is definitely a good option. The only thing I'm unsure about is if convex officially supports better-auth, and they only seem to have one person working on it (who seems to be doing an excellent job), but there's clearly some key man risk here. I'm not sure how complex the underlying convex-better auth integration actually is for the convex team to allow convex users to use both systems, so if this person decided to leave convex, it's not clear how soon maintenance would continue after.
So, if your project has commercial investment behind it, maybe using an established provider like workos or clerk might be a safer long-term choice (at least until there's clarifications on the points above). If this is more of a side project, using the better-auth integration likely wont hurt.
@erquhart Seems to be the one moving the ball forward on the project. Maybe, @erquhart, you could comment on a few of the points discussed?
i succesfully completed a really good comprehensive integration of betterauth into my upcoming app, and then decided to migrate it all to workos, would rather have done that from the start. decided against clerk as i already use it for another project and its not as cost-friendly to new projects unless they're only targeting customers who would be otherwise sufficiently served with betterauth
Nice to have some feedback!Let's say I still have some days to decide which one to pick
@erquhart any comments on this?
@metiuzucc also worth looking on reddit and the #better-auth for people's expereince so far. I've read some after this convo yesterday and there's quite a few people who have encountered bugs/difficulties. I think it's likely due to there being so many edge cases and this integration being relatively new. (This isn't to shame the work done so far in any way, I'm sure it'll only improve over time)
i LOVE the work done so far but its not there yet for my situation to rely on it fully yet but i would use it for b2c applications without hesitancy and backed up my working betterauth implementation for that use case later (and all the breaking change updates were a pain in the butt in a big way, i can't rely on a solution with constantly breaking updates)
Honestly I went with plain convex auth for so many projects, and I really liked the ease of use... I would definitely use it again, just wanted to know if it is going to be deprecated/discontinued... clerk/workos seem over engineered for my use cases now and I really like the fact that I don't need to sign up for third party services