According to a recent LinkedIn post from Replit, the company is highlighting a revamped Free Mode that changes how its Agent can be used under Core and Pro subscriptions. The post suggests that everyday tasks such as chatting, ideation, and building no longer consume credits, with Core subscribers reportedly able to create up to 30 times more and allowances resetting every five hours.
The post indicates these capabilities are powered by OpenAI’s GPT-5.6 Luna and emphasizes new “Conversations” features that keep context from initial ideation through to app creation and across projects. Replit also notes renamed Agent modes, with Economy becoming Power and Power becoming Max, and suggests that users building in Free Mode will receive recommendations to upgrade only when more intensive tasks require it.
Additional features described include Routines in beta, which can schedule tasks from conversations and deliver results back into the same thread, such as daily summaries of Slack, email, or calendar activity. The post also references “Memories,” a capability that appears designed to remember user work preferences, potentially reducing friction for repeat development workflows.
For security-focused users, the LinkedIn post highlights an expansion of black box penetration testing, extending Replit’s existing package and codebase scanning to simulate external attacker behavior. Users are encouraged to run a Level 3 scan from the Security Center, which could appeal to more compliance-conscious or enterprise-adjacent developers seeking stronger application security assurances.
From an investor perspective, these updates suggest Replit is working to increase engagement and usage within existing subscription tiers while adding value through automation, personalization, and security features. If successful, this strategy could improve customer retention, drive upsell from Free Mode to higher-powered Agent tiers, and enhance Replit’s competitive position in the AI-assisted software development and DevSecOps tooling market.