Introducting Telegram plugin for Codex
Back in March I tried the Telegram plugin for Claude Code and liked what it did to my workflow: the session stays on my machine, but I can talk to it from my phone. I wanted the same for Codex. It didn't exist, so I built it seven-ish times
Man, this little side journey was quite nasty in the start, in the middle but at last working in the end
Okay, looks like the problem was me
Astra working on Max fast slowly eats 1% an hour and I can't get it move quota faster, by now if i'd use Chatgpt app it would go -30% easily
I optimized my .md's, global instructions, reset all plugins and skills, as well as project md's
Jev Engineering is what lets one decision layer control hundreds of agent paths without turning the system into chaos.
and up to 193x faster and 444x cheaper in tests.
it’s deciding:
→ which model gets called
→ which tool gets access
→ which result survives
that’s what the video is showing.
one state enters.
Jev turns it into structured decisions.
those decisions fan out across dozens of routes, agents and tools.
then everything collapses back into one verified state.
so instead of:
LLM
→ agent → tool → another LLM → another guess
you get:
state
→ Jev decision layer → parallel routes → execution → verification → next state
that’s the point of Jev Engineering:
more parallelism without giving every branch permission to improvise.
full breakdown in the article below ↓
@RobertJBye I think the bloated nature vs cli is the most concerning one, perhaps there should be a switch to disable all unused plugins and skill and use them per needs basis
Okay, I figured out Astra quota problems
I was so tired to battle burning through another 20X, so decided to get back to Codex CLI and removed all plugins, skills, opencodego router or whatnot, just pure codex with 1M context window and github
Wow, Astra medium ate only 6% quota so far that usually by now should be 30-40% and no work done
It used subagents on Sol High, Terra High-XHigh for code audit and runtime production work; so no light stuff
And speed of work is like I’m using /fast on desktop or even 2x/fast
Perhaps it was bloated app software tech after all and cli is the way as usual
@cherry_mx_reds Ah, this way, I thought they actually select a few; but I noticed how good Opus recently is, so much work done and fast with so little quota spent