nevo-david, solo founder of Postiz (social-media scheduling), says making it agent-ready took revenue from 103k a month in two months, with no VC and no employees. How: expose the product as a skill (an MD file built from the API docs) for personal agents and as an MCP for remote agents like ChatGPT and Claude. Then get discovered via GitHub “awesome” lists (SEO/AEO), ClawHub, and the official ChatGPT/Claude marketplaces; Claude listings (1.2M downloads) lead to Claude citing Postiz. CLI versus MCP: CLIs are lighter, faster, and keep credentials out of the model’s context (env vars or OAuth login); MCPs carry heavy JSON context but are needed for ChatGPT and stateful cases. Build both, and design CLIs with good help and gradual exposure. The bigger picture: agents increasingly pick and embed vendors (you add the credit card later), and consumers make decisions through ChatGPT. Agent-ready products capture new demand that most SaaS still ignores.
Show: langtalks
Concepts: agent-ready-products · agent-security · skill-engineering · agent-workspaces