Replit Introduces Free Mode for AI Agent in Partnership With OpenAI

Replit has rolled out a new Free Mode for its automated coding assistant, significantly lowering the barrier for developers and casual creators using its platform. The update was revealed in a joint announcement by the chief executives of Replit and OpenAI.

Under the new structure, routine operations such as planning projects, chatting with the assistant, and issuing quick generation prompts will no longer rapidly exhaust standard account credits. The system operates on OpenAI's GPT-5.6 Luna model, applying a usage cap that refreshes on a five-hour cycle.

Subscribers on Replit's Core plan will see an expanded capacity under the overhaul. According to the company, these paying users can expect up to 30 times more creation volume alongside roughly 30 hours of monthly conversational agent access.

Unified Workspace

Alongside the tier adjustments, Replit unified its chat interface and software build environment into a single workspace. The integration allows the AI agent to translate ongoing conversation threads directly into active codebase changes without requiring manual prompt transfer.

The update reflects an aggressive push across developer platforms to democratize application building. As frontier labs improve model efficiency, coding platforms are increasingly absorbing baseline inference costs to lock in user mindshare before rivals can establish standard workflows.

By leaning on more cost-effective sub-models like Luna for conversational planning and reserving premium compute for heavy lifting, developer toolmakers can offer virtually continuous agent interactions. Whether this tier expansion converts casual tinkerers into long-term enterprise customers remains the crucial test for Replit's business model.