A bootstrapped team has shipped two open-source products targeting the gap between what AI agents can do and where they’re allowed to exist. KinthAI is a collaboration network where humans and AI agents share group chats with persistent memory, and tAI is a social feed built specifically for AI agents to post and interact, the team announced on DEV Community on April 5.
Both products are built on OpenClaw and support Claude, GPT, Gemini, and any LLM via the open-source plugin on GitHub. KinthAI includes multi-agent group chats (multiple agents collaborating in one conversation), persistent memory across sessions, and an agent marketplace where creators can list and monetize agents at 0% platform fee during beta. tAI takes a different approach: it’s a Markdown-native, API-first social platform where agents post without human puppeteering, all content is publicly indexed, and no login wall sits between the output and the web.
Why It Matters
Every major platform actively filters or bans automated accounts. Reddit, Twitter, and Discord all rate-limit or flag bot behavior. The KinthAI team’s bet is that agent-generated content needs its own infrastructure rather than trying to pass as human on platforms designed to exclude it. The tAI feed gives each agent a persistent public profile where outputs compound over time instead of disappearing after a chat session. The team references Character.AI’s 45 million monthly active users as evidence of demand for AI interaction, then argues the current model is broken: no memory, no creator revenue, no multi-agent capability, and heavy content restrictions.
No VC funding disclosed. The stack runs on Node.js, React, PostgreSQL, and OpenClaw, deployed on bare metal.