Replit Agent

Summary

Replit Agent is an AI assistant that helps developers build, debug, and deploy code projects directly within the Replit IDE.

Release history

1. snapshot-20260829 seen Aug 29, 2026 · issue 010

   Replit Agent gains intelligent model routing, Growth Skills for customer acquisition, and custom OAuth connectors for Enterprise.

   └──▷ HOW TO FIND IT
   Set up Gmail connector with your own OAuth app so outbound emails route through your organization's verified sender identity.
   📍In the Enterprise Workspace, go to Integrations › Connectors › Gmail, select 'Use custom configurations', and enter your Client ID, Client Secret, and desired scopes.
   Run a Growth Skills motion to find early customers for your app using a connected partner integration.
   📍Open Growth Skills in the Agent sidebar, select a partner (e.g. Apollo or PostHog), and run a motion against your deployed app.
   • ›Adds 'Auto' mode for intelligent model routing, which automatically selects the model and effort level best suited to each part of a task, with the option to override to a specific model.
   • ›Intelligent Model Routing is now the default Agent mode for Enterprise Workspaces, with account admins able to control model availability per Workspace and toggle Auto mode on or off.
   • ›The Enterprise Agent mode menu now offers 'Power' and 'Max' tiers (replacing the former Lite option).
   • ›Adds Growth Skills, a new capability that runs Agent-powered go-to-market motions to find an app's first customers, with integrations for Apollo, Clay, SideShift, RevenueCat, Stripe, and PostHog.
   • ›Enterprise organizations can now configure custom OAuth credentials — including Client ID, Client Secret, and scopes — for connectors (e.g. Gmail) instead of using Replit's defaults, via 'Use custom configurations'.
   └──▷ BREAKING ON UPGRADE

   • !The Enterprise Agent mode menu no longer includes the Lite option; only Power and Max are available.
2. changelog-20260828-fda5f4d6 Aug 28, 2026 · issue 009

   Replit Agent adds Free Mode, Conversations, Routines, Steer, GitHub Skill import, enterprise model controls, and Level 3 black-box pen tests

   └──▷ HOW TO FIND IT
   Adjust whether mid-turn messages steer Agent immediately or queue for after the current turn, configured separately for Conversations and Projects.
   📍In the Replit UI, go to Settings › Personalization and set the Steer or Queue preference independently for Conversations and for Projects.
   As an Enterprise admin, restrict which AI model providers and specific models are available to a given Workspace to meet organizational policy.
   📍In the Replit Enterprise admin console, go to Advanced settings, set a per-provider policy to 'Enable selected', then choose the specific models permitted for the target Workspace.
   • ›Adds Settings > Personalization config for choosing Steer or Queue behavior separately for Conversations and Projects; Cmd+Enter (Mac) / Ctrl+Enter (Windows/Linux) toggles the opposite behavior for a single message.
   • ›Enterprise admins can set per-provider model policies (Enable all, Enable selected, or Disabled) and control which models are available in individual Workspaces via Advanced settings.
   • ›New Free Mode lets users chat, explore, and build without consuming Power Mode or Max Mode credits; remaining allowance and reset timing visible in Settings > Usage.
   • ›New Conversations surface lets users ask questions, attach context, or describe an outcome privately before promoting to a full Project, with all context and files carried forward.
   • ›New Routines feature lets users schedule recurring Agent work inside a Conversation, with each result returned to the same Conversation for review; runs consume Power Mode or Max Mode budget per run.
   +3 moreshow less

   • ›Steer capability lets users send follow-up context to Agent mid-turn rather than queuing it after the current turn completes.
   • ›Supports importing Skills from a public GitHub repository, folder, or file URL with a preview step before adding; saved Project Skills now support Required, Available, or No access member access levels.
   • ›New Level 3 security scan combines dependency/package checks, Agent static analysis of source code, and an external black-box test of the live Preview, with the source-code review and live test running in parallel.
3. launch-20260826-e36a5581 Aug 26, 2026 · issue 009

   Replit Agent adds Intelligent Model Routing, automatically selecting the best AI model per task while cutting costs 65% vs. Max Mode.

   • ›Introduces Intelligent Model Routing, which dynamically selects the optimal AI model for each task — balancing quality, speed, and cost — without requiring the user to choose a model manually.
   • ›Delivers the same output quality at 65% lower cost compared to the previous Max Mode, expanding what Free Mode users can build.
   • ›Notifies users when their work escalates from Free Mode to higher-powered (cost-incurring) modes, with the option to override and stay in Free Mode.
   • ›Core and Pro users retain the ability to manually select specific models instead of using automatic routing.
   • ›Enterprise administrators can define an approved model set per workspace; Intelligent Model Routing then selects automatically only from within that approved set, keeping AI usage within company policy.
4. docs update Aug 23, 2026 · issue 005

   Replit Agent gains enterprise model controls, SSO via Clerk/SAML, a Project Security Center, agent-driven security scans, and black-box pen testing.

   └──▷ HOW TO FIND IT
   Set up SAML SSO for your Replit Enterprise org so all members authenticate through your corporate IdP without managing separate credentials.
   📍In your Replit Enterprise workspace, go to Settings › Identity and Access Management › SAML SSO, enter your IdP metadata URL or XML, configure email domain claims, then save to activate.
   Run a black-box pen test against a deployed Replit App to catch runtime vulnerabilities that static analysis misses.
   📍In the Project Security Center, select 'Level 3 - Black-box pen test', start the scan, then review browser and network interaction findings in the findings panel.
   • ›New Enterprise model controls let admins choose which Agent model providers and models are available in each Enterprise Workspace.
   • ›New SSO integration via the Clerk dashboard lets app users sign in with company identity providers (Okta, Microsoft Entra ID) plus multi-factor authentication and session controls.
   • ›New SAML SSO setup for Replit Enterprise Organizations is now configurable directly from workspace settings, including IdP configuration and email domain claims.
   • ›New Project Security Center lets developers choose a security scan, review findings, and fix vulnerabilities before publishing a Replit App.
   • ›New agent security scans enable source-code security review, finding assessment, Agent-assisted remediation, and fix verification in a single workflow.
   +2 moreshow less

   • ›New black-box pen tests (Level 3 security scan) combine source-code review with live browser and network interaction testing against the running app.
   • ›New Steer and Queue controls let users redirect or schedule follow-up messages to Agent mid-task, with Conversation and Project-level preferences.
5. snapshot-20260823 seen Aug 23, 2026 · issue 005

   Replit Agent gains Free Mode, Routines scheduling, mid-run steering, GitHub Skill import, enterprise model controls, and Level 3 black-box pen testing.

   └──▷ HOW TO FIND IT
   Switch mid-run message handling to Queue for Projects when you want Agent to finish its current turn before acting on your next message.
   📍In the Replit console, go to Settings › Personalization and set the Projects message mode to 'Queue'. To override for a single message while Agent is running, press Ctrl+Enter (Windows/Linux) or Cmd+Enter (Mac).
   Run a Level 3 black-box pen test to combine dependency checks, static analysis, and a live external scan of your deployed Preview in parallel.
   📍In your Replit Workspace, open the Security panel and select 'Level 3' scan. Replit will run the dependency and package checks, Agent static analysis of source code, and an external black-box test of the live Preview concurrently.
   Restrict available AI models across your organisation by disabling a provider for all Workspaces, then selectively re-enabling approved models per Workspace.
   📍In the Enterprise admin console, go to Settings › Model Controls, set the provider policy to 'Disabled' for the target provider, then open an individual Workspace and set its model access to 'Enable selected', choosing only the approved models.
   • ›Adds Free Mode to the Agent modes panel, letting users chat, explore, and build without consuming Power Mode or Max Mode credits; model is selected automatically via Auto and remaining allowance is visible at Settings > Usage.
   • ›Adds Routines to schedule and repeat work inside a Conversation, with each run returning results to the same Conversation; active Routines, schedules, and run history are reviewable in one place (runs consume Power Mode or Max Mode budget).
   • ›Adds Steer as the default mid-run message mode for Conversations and Projects, letting users inject new context into an active Agent turn; preference switchable to Queue per context at Settings > Personalization, with Cmd+Enter (Mac) or Ctrl+Enter (Windows/Linux) toggling the opposite behavior for a single message.
   • ›Enables importing Skills from a public GitHub repository, folder, or file URL with a preview step before adding; saved Project Skills now support per-member access levels of Required, Available, or No access.
   • ›Adds enterprise-level Agent model controls to set an account-wide policy per model provider and then choose model availability per Workspace using Enable all, Enable selected, or Disabled.
   +2 moreshow less

   • ›Adds a Level 3 security scan that combines dependency/package checks, Agent static analysis of source code, and an external black-box test of the live Preview, running the source-code review and external test in parallel.
   • ›Adds Conversations as a new starting point inside a Workspace, letting users ask questions, attach context, run a Routine, or describe an outcome before escalating to file-editing work.
6. snapshot-20260820 seen Aug 20, 2026 · issue 002

   Replit adds workspace region selection, native MCP integration, Admin API beta, inline charts, and shareable workspace skills.

   • ›New Admin API (beta) lets Enterprise admins create API keys from the Developer tab in Settings to view workspaces, members, user groups, projects, deployments, usage over custom date ranges, and view/edit budgets for workspaces, user groups, and individual members.
   • ›Adds native MCP Server support so any compatible MCP client (ChatGPT, Claude, Slack, or custom) can create, find, inspect, update, and publish Replit apps, including forking an app into a private editable copy.
   • ›Pro users can now select a workspace region (North America, Europe (EU), or Asia) at workspace creation time, scoping new projects and workspace resources to that region.
   • ›Adds Cal.com, Airtable, and Rootly Docs as connectable MCP servers from within Agent.
   • ›Agent now renders answers to data questions as inline charts directly inside the conversation.
   +2 moreshow less

   • ›Workspace members can create, privatize, or share skills; shared skills default to off for other members; admins can manage all workspace skills and enforce required skills across a workspace via a redesigned Skills Library.
   • ›Agent can now migrate an eligible Replit Auth app to Clerk Auth, moving account data before updating sign-in code so existing users retain access.