Activity
Mon
Wed
Fri
Sun
Jun
Jul
Aug
Sep
Oct
Nov
Dec
Jan
Feb
Mar
Apr
What is this?
Less
More

Memberships

Claude Code Architects

1.1k members β€’ Free

Chase AI Community

59.2k members β€’ Free

AI Automation Society Plus

3.5k members β€’ $99/month

Maker School

2.2k members β€’ $184/m

AI Automation Society

347.3k members β€’ Free

Web Agency Accelerator (FREE)

14.9k members β€’ Free

AI Automation (A-Z)

153.2k members β€’ Free

AI Operators Club

10.5k members β€’ Free

The AI Advantage

120.4k members β€’ Free

46 contributions to AI Automation Society
🚨 URGENT: Claude Code is silently burning my session usage. Has anyone found a fix?
I’ve got a client call today and I’m effectively blocked because Claude Code is unusable right now. I’m on Max 5x. As you can see in the attached GIF, my current session usage was at 49% and a slow silent consumption keeps going on. So the issue doesn’t look like normal weekly exhaustion. It looks like something inside the current Claude Code session is burning context extremely fast. I’ve disconnected MCP servers, tried reducing background activity and waited for my weekly limits to reset, but the same issue keeps coming back. Even when I’m barely doing anything, the session usage climbs very quickly. I’m trying to work out whether this is caused by bloated claude.md files, duplicated project context, hidden MCP activity, background indexing, prompt injection or something else repeatedly loading into context. Has anyone built a practical way to audit what Claude Code is actually consuming? I’m not looking for generic advice like β€œwait for your limits to reset” or β€œupgrade the plan”. I need to understand what is consuming the tokens so I can fix the setup and get back to client work. Any transparent breakdown, diagnostic approach or tool/plugin idea would be really appreciated.
🚨 URGENT: Claude Code is silently burning my session usage. Has anyone found a fix?
0 likes β€’ 32m
@Tsvetomir Krumov every time you notice an unusual consumption run an audit in either Codex or Claude and ask to evaluate where are your tokens going. Use the input from people here just to polish your approach. This is what happened to me. Thank you again to the community.
0 likes β€’ 22m
@Imtiaz Hasan I’m only actively managing one real client deployment right now. The issue wasn’t that I had many active deployments. It was more that I have a habit of creating a new workspace for every project, often with its own CLAUDE.md, notes, and context. Combined with persistent memory/hooks, that can become risky because background observers may treat every active workspace/session as something to watch and summarise. So my takeaway is exactly what you said: I need an operating checklist.
This cheat sheet lives on my wall now 😍
Honestly I kept forgetting the folder structure, the hook events, where mcp.json goes Finally someone laid it all out on one page Saved it, printed it, done If you're building with Claude Code and don't have this pinned somewhere you're making it harder than it needs to be :)
This cheat sheet lives on my wall now 😍
0 likes β€’ 18d
thanks @Nigel Vargas enjoy itπŸ‘Œ
1 like β€’ 17d
@Just Ken Ha good catch Ken, eagle eyes definitely not a test... just proof it was made at 2am πŸ˜…
GOOD DAY FOR EVERYONEπŸ€—
Hope everyone’s having a productive day πŸ‘Š What’s one thing you’re working on right now?
1 like β€’ 18d
@Esayas Tesfaye Thanks man! The timing piece is where it gets interesting.. I'm using signal scoring to decide whether a lead gets a nudge or gets left to cool. Things like email opens, link clicks, days since last touch. Early days but already seeing better reply rates than just blasting follow-ups on a fixed schedule. What signals are you using to qualify your leads on the auto-reply side?
0 likes β€’ 17d
@Esayas Tesfaye email opens and link clicks in the last 48 hours.. matters way more than budget or location since those are static. Behavioral data tells you when to reach out, which is the whole game with follow-ups 😁
7-Day Challenge Meet the Builders
I've started working my way through the 7-Day Challenge in the Classroom - Just got a first draft of a landing page re-design done which is amazing! I had what seemed like every issue in the "book" come up... I didn't have Git installed, then didn't have node installed, then there were some issues with VS Code and Claude working together nicely... but I'm learning that the way to be successful is to keep bashing my head against every challenge until the challenge gives way! QUESTION: There's a step for Day 1 that says: Meet the Builders Community Introduce yourself in the community thread. Share what you're hoping to build, your experience level, and what brought you here. You'll see other builders doing the same β€” these are the people you'll be sharing your work with over the next 7 days. Is the Meet the Builders Community a separate thing in this group? Or, does that just mean to post to the General Discussion?
1 like β€’ 18d
@Mark Lyne Medicare/health insurance is a great niche for this. Those agencies live and die by lead capture so a solid landing page is a real deliverable, not just practice. Keep iterating on it, that could easily become a case study you use to land more clients in that space
0 likes β€’ 18d
@Mark Lyne That's the move. Two sellable services from one challenge is solid
πŸš€New Video: Claude Just Told Us to Stop Using Their Best Model
Anthropic just released the advisor strategy, and it's a big deal for anyone building with Claude. Instead of running every task through Opus, you can pair it as an advisor with a cheaper model like Sonnet or Haiku doing the actual work. In this video I built a dashboard to test how it performs across easy, medium, and hard prompts, and I also show you how to use the same idea inside Claude Code with opus plan mode to stretch your session limit further. GITHUB REPO
0 likes β€’ 18d
The opus plan trick in Claude Code is a game changer. I've been burning through my session limit way too fast running everything on Opus. Switching to plan mode for the thinking and letting Sonnet handle execution is basically the advisor strategy but for your wallet. Appreciate you breaking down the API side too, that cost calculator demo makes the ROI impossible to ignore
1-10 of 46
@david-barroso-2714
Turning simple ideas into systems that compound over time

Online now
Joined Apr 3, 2026
ENFJ
London
Powered by