Andric

Indie software dev and designer. On a quest to build software that's pragmatic, ergonomic, and collaborative.

[Andric @astralwave]

@kartanis @thsottiaux +1

It’s like having a loadout in RPG games, would be a super useful quality of life feature

[Andric @astralwave]

@karrisaarinen @linear friendlier

Would love to try this!

Just onboarded into Linear recently because I needed more structure for tasks, but wasn’t pleased with the lack of good docs functionality

[Andric @astralwave]

Re: ID mapping, you mean to store it in the Convex DB and have an index to query by the client-generated IDs?

I'm not so much worried about rewrites (agents can help)

But the ergonomics and economics of querying through Powersync instead, tradeoffs in terms of DB query and mutation expressiveness, and reliability/latency (with one additional service in the hot path.

[Andric @astralwave]

@k081e @0xFotex @odysseus0z @instant_db @ElectricSQL I need to look into Powersync again now that I'm on Convex!

What would you say are the rough edges or tradeoffs when it comes to adding an additional sync service to the mix?

Any chance you'd also offer a lower cloud tier between the $0 and $49 tiers or a PAYG tier?

[Andric @astralwave]

@k081e @0xFotex @odysseus0z @instant_db @ElectricSQL Wait, Powersync supports Convex? I always thought it was just for Postgres

[Andric @astralwave]

Convex doesn’t have a database in the browser (at least not yet).

You can get sync and optimistic updates, but you won’t get optimistic creates or an instant boot since you still need to go through the network for creates and queries.

You can prewarm or cache some queries and it’ll speed navigations up, but it’s not the same.

[Andric @astralwave]

@jamwt @thekitze @instant_db @robelestifanos_ Is client side ID generation shipped yet?

Would be nice to be able to do optimistic creates as well.

[Andric @astralwave]

@jxnlco I’ve got a monorepo and a .agents/skills folder that’s not in the root of that monorepo.

The codex app doesn’t show those skills in autocomplete when typing $ or /, only those from the root.

[Andric @astralwave]

Finally found a productive workflow for QA’ing frontends with agents.

Can’t believe it took this long for me to find a working setup.

In Codex & Claude Code, there’s a few ways the agent can help QA your app in a real browser:

1. First-party Chrome extension
2. In-app browser (only on Codex)
3. Playwright MCP
4. Chrome DevTools MCP

The first two sound tempting, because they’re first-party to your agent.

But for manually testing a web app, I think the best option is DevTools MCP.

It not only lets the agent drive the web browser, like all the rest do, but it’s got access to the full suite of dev tools in Chrome, including performance/network tracing, reading console logs, and debugging.

Works super well with Codex /goal: you can ask it to improve perf to hit certain metrics, and it’ll not stop until it hits those metrics.

[Andric @astralwave]

The Codex app browser is a bit of a pain because it’s tied to a Codex session.

I use Agentation because you can click around, copy all the feedback, paste it in a doc, and write a long prompt full of feedback

Then I can either parcel out parts of that to Claude Code sessions if it’s tricky, or send it off as a goal to Codex.

I’ve found that Codex isn’t that good at visual UI work compared to Claude Code

[Andric @astralwave]

Addicted to telling Codex or Claude Code to “get a subagent to do [task]”, especially since it keeps the main thread clean and you can continue chatting in it.

Makes me wish regular ChatGPT and Claude had this.

This pattern is helpful for ideation/research/planning tasks too!

[Andric @astralwave]

@thsottiaux Make the in-app browser work with authenticated apps. I’ve got an app with Clerk auth and the in-app browser doesn’t seem to work with it.

[Andric @astralwave]

@thsottiaux Let me use a skill as custom instructions for commit message conventions.

[Andric @astralwave]

@ajambrosino @OpenAIDevs Add the ability to mention files and trigger skills please! 🙏

[Andric @astralwave]

@kylemathews Works with Codex and Claude Code?

[Andric @astralwave]

@masumparvej_ @dev_alade Haven’t used Hugeicons before, but I just saw this – I’d be interested!

Is this a skill you’re building, or a prompt builder?

[Andric @astralwave]

Not too reliable on the first pass, even with Opus as the default.

I switched to Sonnet and did a second and third pass and it got pretty close.

It definitely can be improved: right now it’s too eager to generate a ton of files.

But I think the approach here of automating design system housekeeping is a promising one, even if there’s still a human-in-the-loop for now.

[Andric @astralwave]

Just tried Claude Design.

Burned through my Pro weekly limit in one session, but boy was I happy.

I don’t think Figma is going to be replaced by this anytime soon. It still works best if you give it hand-designed layouts, color choices, type choices, etc).

That said, the experience is miles ahead of Figma Make.

Only because it makes a serious effort of structuring your design system and it makes creating one from existing artifacts super duper easy (GitHub repo, Figma files, brand assets, copy samples).

I’m an indie maker, and putting together a design system has always been a bit of a chore, but this has legitimately taken a lot of work of my plate.

Now, if only Figma worked this way! Autogenerate or update a design system from the decisions I’ve made in my mockups and my repo.

A boy can dream!

[Andric @astralwave]

@zoink @flippnflops An auto-updating DESIGN.md for any Figma file (referencing linked libraries) would be magical ✨🙏