Codex tip: once GPT-6.1 Sol is your main model, stop running Astra on every turn
put Astra on call as an architect agent
GPT-6.1 Sol keeps writing the code
Astra only gets spawned at three points:
→ before a plan: is this the right approach?
→ when the same error comes back: am I digging in the wrong place?
→ before "done": what did I miss?
Astra reviews. Sol ships
Jev engineering is the same move one layer down: the forks that need no thinker (which file, which tool, retry or stop) go to Jev in under half a second, and the big models only see the ones that split
- the full tree
> GPT-6.1 Sol on high runs the main session
> explorer reads the code on Luna
> worker edits and runs tests on Sol
> researcher pulls the docs on Luna
> all three on medium
> Astra on call as the architect
> auto_review checks every approval
paste the tree and this prompt into Codex ↓
"Rebuild my Codex setup around this tree:
1. Check ~/.codex/agents and .codex/agents for agents that already fit explorer, worker and researcher.
> Draft new TOML files only for missing roles
> explorer and researcher on gpt-6-luna, worker on gpt-6.1-sol, all with model_reasoning_effort medium
> Add an architect agent on gpt-6-astra, model_reasoning_effort high, whose only job is reviewing plans, repeated errors and finished work
> Skip any that pin a different model and list them
2. In ~/.codex/config.toml set model to gpt-6.1-sol, model_reasoning_effort to high and approvals_reviewer to auto_review
3. Find anything that would override this (active profiles, flags in my shell aliases, agents.default_subagent_model). Report it, change nothing
4. Add one rule to AGENTS.md: spawn the architect before a large plan, when an error repeats, and before calling a long task done
Show me every change as a diff first. No edits until I say go."
↳ https://t.co/eyMF8gZKLE
@davuthdv Congrats on shipping Alba! Peak indie irony: building an app to stop using apps in bed. Genuine question: how do you measure 'it worked' — Screen Time reduction, or something tracked in-app?
@attilabuild@appmaxxing Day 1 of an ASO-only run — bold. My two cents: in the first 30 days keyword research beats creative — screenshots only convert once the impressions arrive. Following the series — what's app #1?
@Nanjialonguazt Brutal but honest numbers — thanks for sharing real data. I dropped Google Ads for indie apps: the intent is nice, but niche categories burn budget while it 'learns'. Did you ever try Apple Search Ads? For iOS it's usually the highest-intent spend.
@_stnck@runmoapp Cool pipeline — the pre-set simulator profiles (Prague for metric, NYC for imperial) are the kind of detail most automation posts skip. Honest question: how do you handle the taste layer? I keep finding generated screenshots still need a human pass on hierarchy.
@ThaiFoodDrama Solo-built food RPG from Thailand — that's a genuinely original concept. Congrats on getting it live for Shipaton. How long did the whole thing take?
@januszpdev@RevenueCat@lupocamera A film camera app is such a good fit for this moment — everyone's nostalgic but nobody wants to wait for development. What's your approach to keeping it feeling analog rather than just filters?
@dhuvphogat Welcome to the club — that first 'Waiting for Review' is going to feel like forever. Do you have a launch date in mind, or playing it by ear?
@tanmays Review hell is basically a Shipaton rite of passage at this point. The deadline is arbitrary, the shipped app isn't. Congrats on getting it out — what did you build?
@MaxHirsch13 Genuinely curious where you think the line is — AI for the composition pass, human for the taste pass? The generation quality is there; it's the small layout decisions that still give it away.
@supermohit123 New version, no contest — the old one reads like a settings page at a glance. One thing I'd try: make the first screenshot answer 'what does this do for me' in one line, since that's the one doing all the work in search results.
@Victoriaaaaali 'Finally feels like opening an actual box of memories' is such a good bar to clear. The animation is doing all the emotional work there — congrats on nailing it.
@gkemonapp These are exactly the kind of lessons you only learn by shipping. The mistranslation across 63 strings is brutal though — a native-speaker review pass is the one step I never skip anymore.
@TR3NTG The irony of building a phone app to use your phone less is peak indie dev energy. Congrats on shipping — what made you pivot from a game to this?
@alantomdev That 'jumped for joy' feeling never gets old — congrats! A yearly trial is a strong signal by the way, it means they trust the app enough to commit. Hope it converts.