@convex patterns plugin for claude code - prevents the errors I kept hitting.. problems this solves:
- "Too many bytes read" (16MB limit)
- Auto-suggests .take() vs .collect()
- ReturnsValidationError in prod
- Reminds to update ALL validators
- Duplicate users across dev/prod
- Stable Key auth pattern
- TypeScript circular errors (TS7022)
- Known fixes documented
- Forgot --prod flag, data in wrong env
- Deployment checklist patterns
Real example that triggered this:
- User had 500+ saved items
- Query used .collect() instead of .take(50)
- Hit 16MB limit, Server Error
Also prevents:
- Missing validators leading to runtime crashes
- Wrong indexes causing slow table scans
- Schema changes breaking old queries
- Environment variable mismatches
used /goal to migrate my app from Convex DB to SQLite. Codex took like 8+ hours straight, then needed ~3 more prompts after for fixing/random cleanup ๐
but finally... dashboard loads successfully now
context:
app tracks Skool groups. initially built everything on Convex since wanted reactivity. then the dashboard/analytics part started growing.
I also used Hermes agents to do automated analysis queries directly on the convex db...
BIG mistake ๐ the agents were eating WAY more Convex bandwidth than actual human users. it kept on doing expensive convex operations on the db.
ended up migrating analytics/dashboard stuff to SQLite instead. Codex /goal helped me do this easily.
Higgsfield MCP + Claude Code = AI Ad Agency
Use Higgsfield to generate AI statics & AI UGC...
All without leaving Claude.
00:00 โ Intro
00:55 โ Connect Higgsfield MCP to Claude Code
02:09 โ Build the brand brief with Firecrawl
04:35 โ Generate the hero static with GPT Image 2
09:11 โ Add text overlay to the static ad
10:25 โ Animate the hero shot with Seedance 2.0
12:54 โ Generate the UGC creator with GPT Image 2
14:53 โ Generate UGC video clips with Seedance 2.0
17:25 โ Recap & wrap
$169,947 MRR => ~18,883 members paying just $9/month... they do fb ads at that price point. How is that even profitable? theyโve been doing it since at least Q4 last year and itโs still growing like crazy... Nov last year, they were only 300 members, at about $2k MRR.
AI Video Bootcamp on Skool focused on AI creatives(videos and stuffs). Basically teaching people how to make money with AI-generated ads, social content etc.
yap made it since got curious, want to track actual paid groups growth on specific ai markets/verticals.. built / populate the fb, then have my hermes agent do analysis/evaluation.. e.g. good to see like what actual paid groups started this year -> which ones are killing it and what market/category
@Viswana34226652@onlinedopamine@JulianGoldieSEO own app.. still gated , only my email can access.. really great to see actual skool groups growth across different markets.. loving the data
@Teknium so to trigger this from hermes cli, must use the hermes 'kanban' right? or probably will just hermes agent to do the commands... and purpose of this is really for human visual check?
@sudoingX so you're using the hermes cli as your main harness? since currently using codex cli as my main harness, and wondering if switching is worth it.. i did experience like TUI bugs e.g. copy pasting images and even texts from my environment (windows, ssh on vps running hermes cli)
Skool group just started at 4k USD MMR, from Gambia, West Africa. He just launched it on May 1, 2026.... facebook as his distribution is ๐... he's sharing his AI video workflow. Just curious on how skool groups on the AI space are doing
for created subtask, will they really be visible on the kanban? since i just prompted 'make a video about kanban' at the hermes kanban UI, assigned to default ai model, no workspace, no tenants etc... then watched the card, didn't notice any subagents/subtasks being spawned.. and this is the created video @NousResearch
Hermes Agent now has multi-agent via the Kanban, new in v0.12.0.
Agents claim tasks from a board, work in parallel, and hand off when blocked. You watch progress and unblock from one easy view instead of juggling terminals.
We asked it to plan and make this video about itself: