OpenAI introduced Dots last month as “always-on agents” running in their own cloud computer, but users are reporting task failures and connection issues, dots that disappear, and refusals tied to abuse prevention, per GitHub issues and the OpenAI Developer Community forum. OpenAI did not respond to a request for comment. On social media, some users question whether Dots was rushed in response to Meta Muse and xAI’s Grok Bot; ABI Research argues “AI claws,” always-on agents, will shape enterprise AI adoption.
Competition arrived within weeks. Tawkit, which makes the CopilotKit SDK, released OpenDots, “completely open source,” as a self-hostable template that brings any agent through the AG-UI protocol and uses any OpenAI-compatible model, so “OpenAI may not be doing the billing.” Separately, Prathit Joshi, a growth engineer at Composio in Bangalore, released Open Dot, an open-source desktop app that runs on a user’s own Mac with their own OpenAI key or open models like Kimi, DeepSeek and Qwen through OpenRouter, connecting to Gmail, Calendar, Slack, Notion and GitHub via Composio. Joshi said he built it for use cases closed model companies are not incentivized to support, like allowing open-source models.
Musk said on Tuesday that Grok Bot would become a model router rather than a front-end for one model, using “the best back end model for any given task, including Claude Opus 5.5, MidJourney, Suno and other leading APIs.” The pattern across all three moves: everyone is wrapping increasingly commoditized models in a gatekeeping layer, hoping to build a sustainable business on top.