Fetching from the wire…
01
02
03
04
05
06
07
08
Public story · 2026-08-25 · high
A sales-CRM tool, a dev orchestrator, a usage meter, and a book app all launched inside Claude's ecosystem on August 25 instead of on a homepage.
Why now: All four launches surfaced in the same August 25 Product Hunt and Hacker News Show HN rankings.
Four products launched on August 25 with nothing in common except where they chose to live.
Diet Claude reached #2 with 140 votes on the day's Product Hunt leaderboard. It only tracks how close someone is to hitting Claude's usage limits. A feature that narrow only draws 140 votes if people already spend their workday inside the agent.
Ninjō AI took #10 with 81 votes for AI sales agents that run on Instagram and WhatsApp. Builders manage them through MCP from Claude Code, ChatGPT, and Codex. Coolplugz claimed #7 with 89 votes, a Claude Code orchestrator that pulls context from Jira, GitHub, Notion, and Slack. Coffeetable, the day's top-voted Show HN post, skips a website entirely. It ships as an entry in claude.ai's directory.
A sales CRM, a dev workflow tool, a usage dashboard, and a book discovery app have nothing to do with each other as products. What they share is a decision to let an AI client be the front door instead of building one of their own.
What the leaderboard doesn't say is how many of these four keep users past launch day. Vote counts on Product Hunt and Hacker News measure attention, not whether anyone opens the agent again tomorrow.
Each link below shares sources, entities, or timing with this story.
AgentSky topped Product Hunt's August 3 leaderboard with 220 points; its cloud cloning copies a local Claude Code or Codex agent into the cloud with one prompt.
On August 25, Ninjō AI shipped AI sales agents for Instagram and WhatsApp built and managed through MCP from Claude Code, ChatGPT and Codex; coolplugz shipped a Claude Code orchestrator pulling context from Jira, GitHub, Notion and Slack; Diet Claude shipped a usage meter for...
Unabyss and ZooData beat every app on the leaderboard back to back by selling storage for agents instead of a use case.
Three days running, Product Hunt's leaderboard has favored tools that run agents over agents that do anything themselves.
Five of Product Hunt's ten highest-voted Aug 22 launches manage agents already running on a Mac, echoing the Aug 21 leaderboard.
The 284B-parameter model claims a six-fold agent capability jump over the prior Flash tier while undercutting proprietary pricing by roughly 60 percent.
MindPattern daily
One email a day at 7 AM. Sources and a take on every story. Unsubscribe anytime.