Browser-based IDE with the Replit Agent that scaffolds, edits, and runs apps from prompts. Built-in deployment, hosting, and a database. Closest competitor to Lovable for the prompt-to-app use case.
Replit introduces Cloud Agents for Enterprise that automatically find relevant code, make changes, test them in running apps, and open GitHub or GitLab pull requests. Changes run in isolated cloud sandboxes following team guidelines, enabling non-engineers to contribute while engineers maintain control.
Agent now supports sending messages to other chats and creating new chats without manual copying and pasting. This capability is available to Pro and Enterprise users, streamlining workflows for managing multiple conversations.
Developers can now select GPT-6.1 Terra from the model selector when building with Agent in Power Mode, adding another model option to choose from when configuring their AI assistant.
Agent can now investigate production issues using app logs and metrics, then attempt and validate fixes. Automated triage is available via Monitoring alerts and Auto-fix settings, with optional automatic republishing of validated fixes.
Agent can suggest new skills or improvements to existing ones based on user workflows and conversations, which users review and approve before saving to their workspace skills library.
Users can share entire chats or individual assets with selected teammates or workspaces in read-only mode, including access to development URLs that provide live server access.
Agent can help users incorporate their business through Stripe Atlas directly on Replit, offering a 20% discount on incorporation for users who complete it through the platform.
Eligible top-up customers can keep published apps running without consuming credits up to an overage limit. Free mode remains available with warning emails sent before suspension.
Replit's Agent can now investigate production issues using logs and metrics, then attempt and validate fixes. With Monitoring alerts and Auto-fix enabled, Agent triages issues in the background; validated fixes can optionally auto-republish or await manual review. This shifts on-call burden to AI-assisted automation.
Agent can now suggest new skills or improvements to existing ones based on useful workflows from conversations and projects. Developers review and approve changes before saving to their workspace's skills library, turning ad-hoc solutions into reusable templates.