4 AI launches from August 1 for follow-ups, agents, mobile coding, and selling expertise
Four August 1 launches give creators and online businesses new ways to recover quiet sales threads, package expertise, handle customer questions, and monitor coding agents.
Three launches to test
This edition covers products first listed or materially launched from August 1, 7:45 a.m. through August 2, 7:45 a.m. Eastern Time. The strongest fits are not new models; they are small workflow products for recovering sales conversations, packaging expertise, and handling repetitive customer questions. A fourth launch lead is included at the end, but its public listing is too thin for a full recommendation.
| Tool or platform | What changed | Best fit | First test |
|---|---|---|---|
| NudgeForMe | A follow-up agent scans sent conversations and drafts the next email inside your mailbox. 1 2 | Freelancers, agencies, and small businesses that sell through email | Review 10 quiet threads in draft mode |
| Kopai | A no-code agent builder lets you ground an agent in your documents, publish it to a marketplace, and charge per message. 1 3 | Consultants, educators, and niche experts with repeatable answers | Load one small FAQ set and test citations |
| ElevenAgents | ElevenLabs launched an AI voice-and-chat agent entry on Product Hunt; its platform supports knowledge, procedures, monitoring, and multilingual conversations. 1 4 | Sellers and service businesses with repetitive inbound questions | Build one FAQ agent and keep it in a test channel |
NudgeForMe turns an old inbox into a follow-up queue
The useful detail is where the agent starts: NudgeForMe scans conversations you already sent, looks for unfinished loops such as proposals, client requests, approvals, and support questions, then creates a draft in the original mailbox thread. It supports Gmail, Google Workspace, Outlook, Microsoft 365, Yahoo, iCloud, and other IMAP mailboxes. 2
That fits creators who sell services, sponsorships, retainers, or digital products through one-to-one email. The product's default is draft mode: you can edit, send, delete, or dismiss each suggestion, and the agent says it stops watching a thread after a reply. The free plan is limited to one mailbox and 100 draft credits per month; the site lists Pro at $12 per month or $96 per year with more capacity and optional automatic sending. 2
The limitation is obvious and useful: a drafted follow-up is not a qualified lead. NudgeForMe's page does not provide an independent accuracy or conversion benchmark, so let it surface candidates rather than hand over your tone and judgment. The Gmail and Outlook integrations are described as the deepest native connections; other providers may behave differently. 2
Try this first: connect one mailbox, review 10 old threads, and label each suggestion keep, rewrite, or dismiss. Leave automatic sending off. Start with the free scan, then compare the time saved with the cost of correcting weak drafts.
Kopai packages expertise as a sellable agent
Kopai's August 1 Product Hunt listing describes a marketplace where people can share expertise through AI agents. Its own site says you can describe an agent in plain English, select among GPT, Claude, and Gemini, test it before publishing, and upload PDFs, guides, manuals, and FAQs so answers are grounded in your material and show citations. 1 3
For an educator, consultant, or specialist creator, the interesting change is the distribution layer. The site says each published agent gets a searchable marketplace page, charges users per message rather than by subscription, and can be embedded on your own site. It also shows generated Word, Excel, and PDF files as a supported output pattern. 3
The sales claim needs restraint. The public page does not state the creator revenue split, moderation rules, or how marketplace demand is generated. Treat the marketplace as an acquisition experiment, not as proof that an audience will appear. Your first risk is answer quality: an agent that cites a document incorrectly can damage trust faster than a static FAQ can.
Try this first: upload three to five documents behind one narrow paid answer—such as a course FAQ, onboarding guide, or client checklist. Ask it 10 questions, inspect every citation, and only then decide whether the result is safe to publish. Start building in Kopai.
ElevenAgents moves repetitive customer conversations into a testable workflow
Product Hunt listed ElevenAgents by ElevenLabs on August 1 with the pitch "Scale conversations without scaling your team." The underlying ElevenLabs platform supports both voice and chat agents. Its setup flow lets you upload knowledge, define procedures and escalation paths, monitor conversations, and refine behavior; the company says it supports language detection and switching across 70+ languages. 1 4
That is relevant to a seller or small service business when the same questions arrive repeatedly: delivery status, booking rules, product fit, or a lead's next step. The official page lists integrations including Stripe, Cal.com, Twilio, Zapier, Zendesk, and HubSpot, plus a no-code web setup and API/SDK access. 4
The access condition is less clear than the feature list. The page offers sign-up and a "talk to sales" path but does not publish pricing in the material retrieved here. The platform also makes broad scale and quality claims; those are company statements, not independent evidence. Start with a narrow FAQ and measure wrong answers, handoffs, and latency before putting it in front of customers.
Try this first: write 10 real customer questions and one escalation rule, build a web-based test agent, and have a colleague try to break it. Do not connect payments, telephony, or a live inbox until the agent handles the ordinary cases and knows when to stop. Create an ElevenAgent.
A thinner lead: Port22 puts coding agents on a phone
Product Hunt's August 1 board also lists Port22 as "Claude Code, Codex & more on your phone." That makes it relevant to creators who build small apps or automations and want to check work away from a desk, but the accessible launch listing does not establish pricing, supported phone platforms, security details, or how the desktop connection works. 1
This one belongs in the watch column, not the same recommendation tier as the three entries above. If you build on a Mac, the safe first test is a non-sensitive toy task: leave one short agent run active, verify what the phone can show, and check whether approvals require exposing credentials or source code. Open the Port22 launch page before installing anything.
Where I would start
- Selling services by email: start with NudgeForMe's draft-only scan. It is the smallest reversible test and keeps the send decision with you.
- Selling repeatable expertise: test Kopai on a narrow document set. The citation check matters more than the marketplace listing.
- Handling customer questions: prototype ElevenAgents in a private channel, with one clear escalation rule and no payment access.
- Building apps or automations: treat Port22 as an experiment until its connection and security details are clear.
The common thread is modest: each launch puts an existing bottleneck—follow-up, expertise distribution, customer intake, or agent monitoring—into a narrower product surface. Pick the one attached to a workflow you already run; do not adopt four new systems because they launched on the same day.
Related content
- Sign in to comment.