# Replit

Podium publishes this catalog and sells an agent development environment. Podium chooses the products, criteria, sources and wording, so treat these pages as vendor-authored research. Check important claims against the linked primary sources. The same evidence rules apply to Podium and its competitors, but they do not remove this conflict of interest.

Page: https://adecomparison.com/ade/replit/
Website: https://replit.com/
Research: covered

Method: https://adecomparison.com/method/

Rank by must-haves met, then preferences, then name. Read each finding's checked date. Findings stay published after their scheduled recheck. Checks over six months old may be outdated.

## replit:ACC-01

Finding: Replit offers browser and phone-browser access, native iOS/Android clients, and native desktop apps on Mac, Windows and Debian-based Linux.

Verification: sourced
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation and pinned source inspection only; no hands-on scenario was run.
Other atomic questions remain unestablished: cli, editor_extensions.

Source: https://docs.replit.com/learn/projects-and-artifacts/project-editor
Locator: The Project Editor
Finding: The hosted Project Editor provides Agent chat, project tools, live preview and a task board in the browser.
Retrieved: 2026-11-08

Source: https://docs.replit.com/platforms/mobile-app
Locator: Mobile features; Navigate your project
Finding: Native iOS and Android clients support Agent, Tools and Tasks panes, project continuation and completion or help notifications.
Retrieved: 2026-11-08

Source: https://docs.replit.com/features/platforms/desktop-app.md
Locator: What is the Replit Desktop App; install and sync
Finding: The current first-party desktop guide explicitly offers a native app for Windows, Mac and Debian-based Linux, including Agent, Project Editor and publishing. This establishes desktop clients rather than inferring them from a companion CLI.
Retrieved: 2026-11-13

Source: https://replit.com/usecases/ai-code-editor
Locator: From your browser or mobile phone
Finding: The first-party AI code editor page explicitly offers the full cloud development workspace from a browser on desktop or mobile devices, alongside the native mobile apps.
Retrieved: 2026-11-13

## replit:ACC-02

Finding: The documented development experience runs Replit Agent with selectable modes/models. The checked sources do not establish named third-party coding-harness integrations; building an app that calls a provider or installing arbitrary shell software is separate.

Verification: sourced
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation and pinned source inspection only; no hands-on scenario was run.
Other atomic questions remain unestablished: custom_adapters, session_integration, mixed_harness_operation.

Source: https://docs.replit.com/features/agent/overview
Locator: Agent modes
Finding: Replit Agent has Free, Power and Max modes; Auto routes paid work to models. Model routing does not establish support for a third-party coding harness.
Retrieved: 2026-11-08

Source: https://docs.replit.com/help/agent-and-ai.md
Locator: Can I use my own API keys with Agent?
Finding: The first-party FAQ separates app API credentials from the Agent itself: user provider keys in Secrets power calls made by the built app, while the Agent experience uses Replit-managed model infrastructure rather than the user-key route. Curated in-product model selection is separate from replacing that infrastructure with an independently keyed endpoint. Replit AI billing charges its own plan credits. This is not evidence of a third-party coding harness or a native subscription-login path.
Retrieved: 2026-11-13

## replit:ACC-03

Finding: Replit plan credits pay for Agent. User API keys power calls made by built apps, while the Agent itself remains Replit-managed rather than using the app API keys. A native external coding-agent subscription login remains unestablished.

Verification: sourced
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation and pinned source inspection only; no hands-on scenario was run.
Other atomic questions remain unestablished: local_models, account_switching, vendor_account_requirement, provider_restrictions.

Source: https://docs.replit.com/billing/ai-billing
Locator: Agent billing
Finding: Replit plan credits pay for Agent and other cloud services; paid actions use effort-based pricing after confirmation.
Retrieved: 2026-11-08

Source: https://docs.replit.com/help/agent-and-ai.md
Locator: Can I use my own API keys with Agent?
Finding: The first-party FAQ separates app API credentials from the Agent itself: user provider keys in Secrets power calls made by the built app, while the Agent experience uses Replit-managed model infrastructure rather than the user-key route. Curated in-product model selection is separate from replacing that infrastructure with an independently keyed endpoint. Replit AI billing charges its own plan credits. This is not evidence of a third-party coding harness or a native subscription-login path.
Retrieved: 2026-11-13

## replit:ACC-04

Finding: Start a project by describing an idea in plain language, or bring it in from GitHub, Figma or uploaded files, in the cloud Project Editor.

Verification: sourced
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation evidence only. Remaining atomic questions and enforcement under failure have not been established by a bounded workflow test.

Source: https://docs.replit.com/learn/projects-and-artifacts/project-editor
Locator: Getting started; Import code and design
Finding: Start a project by describing an idea in plain language, or bring it in from GitHub, Figma or uploaded files, in the cloud Project Editor. The page names no templates and does not itself state that no local environment is needed.
Retrieved: 2026-11-08

## replit:PLAN-01

Finding: A main-thread prompt can create titled tasks with descriptions and plans, or users can start a planning thread.

Verification: agent-cross-checked
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation evidence only. Remaining atomic questions and enforcement under failure have not been established by a bounded workflow test.

Source: https://docs.replit.com/core-concepts/agent/task-system
Locator: How background tasks work
Finding: A main-thread prompt can create titled tasks with descriptions and plans, or users can start a planning thread.
Retrieved: 2026-11-08

## replit:PLAN-02

Finding: Agent proposes detailed plans and definitions of done; users inspect View plan, accept tasks or request a revised plan before execution.

Verification: agent-cross-checked
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation evidence only. Remaining atomic questions and enforcement under failure have not been established by a bounded workflow test.

Source: https://docs.replit.com/core-concepts/agent/task-system
Locator: Plan your tasks; Review and accept
Finding: Agent proposes detailed plans and definitions of done; users inspect View plan, accept tasks or request a revised plan before execution.
Retrieved: 2026-11-08

## replit:PLAN-03

Finding: Agent identifies prerequisite tasks and queues dependent work until the prerequisite is applied. The board shows the blocker.

Verification: agent-cross-checked
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation evidence only. Remaining atomic questions and enforcement under failure have not been established by a bounded workflow test.

Source: https://docs.replit.com/learn/build-in-parallel
Locator: Create and approve remaining tasks; Let Agent schedule the work
Finding: Agent identifies prerequisite tasks and queues dependent work until the prerequisite is applied. The board shows the blocker.
Retrieved: 2026-11-08

## replit:PLAN-04

Finding: Each background task inherits the project context and receives its own Agent and project copy.

Verification: agent-cross-checked
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation evidence only. Remaining atomic questions and enforcement under failure have not been established by a bounded workflow test.

Source: https://docs.replit.com/learn/build-in-parallel
Locator: Your app is safe while tasks run
Finding: Each background task inherits the project context and receives its own Agent and project copy.
Retrieved: 2026-11-08

## replit:PLAN-05

Finding: Projects can begin from GitHub, Figma or uploaded files; this establishes input channels, not verified architecture comprehension.

Verification: agent-cross-checked
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation evidence only. Remaining atomic questions and enforcement under failure have not been established by a bounded workflow test.

Source: https://docs.replit.com/learn/projects-and-artifacts/project-editor
Locator: Import code and design
Finding: Projects can begin from GitHub, Figma or uploaded files; this establishes input channels, not verified architecture comprehension.
Retrieved: 2026-11-08

## replit:PLAN-06

Finding: Native tasks have Drafts, Active, Ready and Done states, a board, separate threads and explicit acceptance. Draft planning sessions can be archived and restored; cancelling building work discards it.

Verification: agent-cross-checked
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation evidence only. Remaining atomic questions and enforcement under failure have not been established by a bounded workflow test.

Source: https://docs.replit.com/core-concepts/agent/task-system
Locator: What is the task system; Frequently asked questions
Finding: Native tasks have Drafts, Active, Ready and Done states, a board, separate threads and explicit acceptance. Draft planning sessions can be archived and restored; cancelling building work discards it.
Retrieved: 2026-11-08

## replit:PLAN-07

Finding: Agent can split a request into tasks and put them on the human-visible task board.

Verification: agent-cross-checked
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation evidence only. Remaining atomic questions and enforcement under failure have not been established by a bounded workflow test.

Source: https://docs.replit.com/core-concepts/agent/task-system
Locator: What is the task system
Finding: Agent can split a request into tasks and put them on the human-visible task board.
Retrieved: 2026-11-08

## replit:COORD-01

Finding: Core supports one active background task, Pro up to ten and Enterprise up to sixty-four. Additional tasks queue.

Verification: agent-cross-checked
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Core 1; Pro 10; Enterprise 64; Starter no background tasks
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation evidence only. Remaining atomic questions and enforcement under failure have not been established by a bounded workflow test.

Source: https://docs.replit.com/core-concepts/agent/task-system
Locator: Tasks start running; Availability
Finding: Core supports one active background task, Pro up to ten and Enterprise up to sixty-four. Additional tasks queue.
Retrieved: 2026-11-08

## replit:COORD-02

Finding: Agent schedules independent tasks when capacity exists and holds dependent tasks until prerequisites are applied.

Verification: agent-cross-checked
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation evidence only. Remaining atomic questions and enforcement under failure have not been established by a bounded workflow test.

Source: https://docs.replit.com/learn/build-in-parallel
Locator: Let Agent schedule the work
Finding: Agent schedules independent tasks when capacity exists and holds dependent tasks until prerequisites are applied.
Retrieved: 2026-11-08

## replit:COORD-03

Finding: Replit Agent decomposes a request into background Agent tasks with separate threads and returns work logs, tests and previews to the main-version review flow.

Verification: agent-cross-checked
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation evidence only. Remaining atomic questions and enforcement under failure have not been established by a bounded workflow test.

Source: https://docs.replit.com/core-concepts/agent/task-system
Locator: What is the task system; Review and apply
Finding: Replit Agent decomposes a request into background Agent tasks with separate threads and returns work logs, tests and previews to the main-version review flow.
Retrieved: 2026-11-08

## replit:COORD-04

Finding: Paid-plan chat Routines repeat hourly, daily or weekly without a new request. They run in Power or Max, queue behind other chat work and do not schedule publishing.

Verification: agent-cross-checked
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Core, Pro or Enterprise
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation evidence only. Remaining atomic questions and enforcement under failure have not been established by a bounded workflow test.

Source: https://docs.replit.com/features/agent/automations
Locator: How Routines run; Availability and limits
Finding: Paid-plan chat Routines repeat hourly, daily or weekly without a new request. They run in Power or Max, queue behind other chat work and do not schedule publishing.
Retrieved: 2026-11-08

## replit:COORD-05

Finding: Cancel stops building and discards its work. The main thread and each task thread permit follow-up direction.

Verification: agent-cross-checked
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation evidence only. Remaining atomic questions and enforcement under failure have not been established by a bounded workflow test.

Source: https://docs.replit.com/core-concepts/agent/task-system
Locator: Frequently asked questions
Finding: Cancel stops building and discards its work. The main thread and each task thread permit follow-up direction.
Retrieved: 2026-11-08

## replit:COORD-06

Finding: A project board displays Drafts, Active, Ready and Done tasks alongside task conversations. Fleet-scale navigation has not been tested.

Verification: agent-cross-checked
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation evidence only. Remaining atomic questions and enforcement under failure have not been established by a bounded workflow test.

Source: https://docs.replit.com/core-concepts/agent/task-system
Locator: Board view
Finding: A project board displays Drafts, Active, Ready and Done tasks alongside task conversations. Fleet-scale navigation has not been tested.
Retrieved: 2026-11-08

## replit:COORD-07

Finding: Users start with ordinary chat, refine plans and let Agent create task lists before approving work.

Verification: agent-cross-checked
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation evidence only. Remaining atomic questions and enforcement under failure have not been established by a bounded workflow test.

Source: https://docs.replit.com/features/agent/overview
Locator: Plan mode; What you can build
Finding: Users start with ordinary chat, refine plans and let Agent create task lists before approving work.
Retrieved: 2026-11-08

## replit:RUN-01

Finding: Replit provides managed cloud development, and the current desktop page also offers local Agent code/build execution on supported Mac/Windows devices. Full BYOC Agent execution is private preview, while published-app exports do not move development.

Verification: sourced
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation and pinned source inspection only; no hands-on scenario was run.
Other atomic questions remain unestablished: self_hosted, hybrid, provisioning, payer, resources, region, expiry, other_device_continuation.

Source: https://docs.replit.com/learn/projects-and-artifacts/project-editor
Locator: Frequently asked questions
Finding: The cloud Project Editor saves files, history and the running app when the browser closes. This documents cloud persistence, not an unlimited Agent process lifetime.
Retrieved: 2026-11-08

Source: https://replit.com/desktop
Locator: Build locally; Isolate builds in local sandboxes
Finding: The current desktop product page explicitly says Agent can build and run apps using the local environment and selected local folders. On supported Windows and macOS devices it names Microsoft Execution Containers for code/build-tool isolation. This does not establish a Linux local-execution sandbox.
Retrieved: 2026-11-13

Source: https://docs.replit.com/teams/bring-your-own-cloud.md
Locator: Compare deployment options; Full BYOC warning and availability
Finding: Full BYOC includes Agent/development execution in a customer tenant but is explicitly private preview, with eligibility/configurations confirmed through Enterprise onboarding. Published-app BYOC and Image Exports are early access and do not move the development/Agent control plane. Even Full BYOC leaves the control plane external and explicitly is not an entirely self-hosted or air-gapped Replit installation.
Retrieved: 2026-11-13

Source: https://docs.replit.com/build/code/working-with-code.md
Locator: Sandbox architecture; draft PR and review workflow
Finding: The current repository-contribution guide explicitly says the development sandbox runs in a microVM. Replit owns the contribution workflow: it creates a branch, commits/pushes changes, opens a GitHub draft PR or GitLab MR, and updates the request after review.
Retrieved: 2026-11-13

## replit:RUN-02

Finding: Replit repository development sandboxes explicitly use microVMs. Its current local desktop offering names Microsoft Execution Containers on supported Mac/Windows devices; background task copies are not established as Git worktrees.

Verification: sourced
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation and pinned source inspection only; no hands-on scenario was run.
Other atomic questions remain unestablished: branch, worktree, os_sandbox, filesystem, process, network, secrets, security_boundary.

Source: https://docs.replit.com/core-concepts/agent/task-system
Locator: Tasks start running
Finding: Background tasks run in isolated project copies and do not change the main version until applied. The docs do not name these copies as containers, VMs or worktrees.
Retrieved: 2026-11-08

Source: https://replit.com/desktop
Locator: Build locally; Isolate builds in local sandboxes
Finding: The current desktop product page explicitly says Agent can build and run apps using the local environment and selected local folders. On supported Windows and macOS devices it names Microsoft Execution Containers for code/build-tool isolation. This does not establish a Linux local-execution sandbox.
Retrieved: 2026-11-13

Source: https://docs.replit.com/build/code/working-with-code.md
Locator: Sandbox architecture; draft PR and review workflow
Finding: The current repository-contribution guide explicitly says the development sandbox runs in a microVM. Replit owns the contribution workflow: it creates a branch, commits/pushes changes, opens a GitHub draft PR or GitLab MR, and updates the request after review.
Retrieved: 2026-11-13

## replit:RUN-03

Finding: Files, history and running app are preserved in the cloud for a later visit. No worker-crash recovery test was performed.

Verification: agent-cross-checked
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation evidence only. Remaining atomic questions and enforcement under failure have not been established by a bounded workflow test.

Source: https://docs.replit.com/learn/projects-and-artifacts/project-editor
Locator: What happens when I close my browser?
Finding: Files, history and running app are preserved in the cloud for a later visit. No worker-crash recovery test was performed.
Retrieved: 2026-11-08

## replit:RUN-04

Finding: Agent sets up the project and infrastructure. Specific dependency-cache, port-collision and clean-remote-host behavior remains untested.

Verification: agent-cross-checked
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation evidence only. Remaining atomic questions and enforcement under failure have not been established by a bounded workflow test.

Source: https://docs.replit.com/features/agent/overview
Locator: What is Replit Agent?
Finding: Agent sets up the project and infrastructure. Specific dependency-cache, port-collision and clean-remote-host behavior remains untested.
Retrieved: 2026-11-08

## replit:RUN-05

Finding: Core and Pro Routines have per-run budgets checked after each step; an in-progress step can complete beyond the budget. Enterprise lacks the per-run budget.

Verification: agent-cross-checked
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation evidence only. Remaining atomic questions and enforcement under failure have not been established by a bounded workflow test.

Source: https://docs.replit.com/features/agent/automations
Locator: Control spending
Finding: Core and Pro Routines have per-run budgets checked after each step; an in-progress step can complete beyond the budget. Enterprise lacks the per-run budget.
Retrieved: 2026-11-08

## replit:RUN-06

Finding: Finished tasks expose a work log, test results and preview; active threads and board cards show task status.

Verification: agent-cross-checked
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation evidence only. Remaining atomic questions and enforcement under failure have not been established by a bounded workflow test.

Source: https://docs.replit.com/core-concepts/agent/task-system
Locator: Review and apply
Finding: Finished tasks expose a work log, test results and preview; active threads and board cards show task status.
Retrieved: 2026-11-08

## replit:RUN-07

Finding: Agent billing is effort-based and uses plan credits shared with cloud services. A controlled reconciliation against provider billing was not performed.

Verification: agent-cross-checked
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation evidence only. Remaining atomic questions and enforcement under failure have not been established by a bounded workflow test.

Source: https://docs.replit.com/billing/ai-billing
Locator: Viewing Agent usage
Finding: Agent billing is effort-based and uses plan credits shared with cloud services. A controlled reconciliation against provider billing was not performed.
Retrieved: 2026-11-08

## replit:RUN-08

Finding: Plan-based task limits cause additional background tasks to queue until slots are available. This is not evidence of per-host CPU or RAM scheduling.

Verification: agent-cross-checked
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation evidence only. Remaining atomic questions and enforcement under failure have not been established by a bounded workflow test.

Source: https://docs.replit.com/core-concepts/agent/task-system
Locator: Tasks start running
Finding: Plan-based task limits cause additional background tasks to queue until slots are available. This is not evidence of per-host CPU or RAM scheduling.
Retrieved: 2026-11-08

## replit:TEAM-01

Finding: Invited teammates work on one project through separate Agent threads and shared task boards. Multiple humans steering the same active Agent conversation or work item remains unestablished.

Verification: sourced
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation and pinned source inspection only; no hands-on scenario was run.
Other atomic questions remain unestablished: discussion, comments, presence, ordering, conflicts, attribution, owner_online.

Source: https://docs.replit.com/features/collaboration/team-workspaces
Locator: Personal vs. team workspaces
Finding: Team workspaces share projects with members; Core includes five collaboration seats and Pro fifteen. Simultaneous co-prompt ordering is not established by this source.
Retrieved: 2026-11-08

Source: https://docs.replit.com/build/invite-teammates.md
Locator: Build with teammates; independent threads and shared board
Finding: Distinct invited teammates can start their own Agent threads and tasks in the same project, share the board, view history and review/apply ready work. The guide emphasizes separate threads; it does not establish multiple humans sending instructions into the same active Agent conversation/work item.
Retrieved: 2026-11-13

## replit:TEAM-02

Finding: Personal workspaces have an admin and project guests; team workspaces add members with access to all workspace projects. Admins can invite and remove members. SAML SSO requires Enterprise and organization/identity-provider admin access. SSO does not automatically invite users; invitations or SCIM manage membership.

Verification: agent-cross-checked
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Enterprise for SAML SSO
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation evidence only. Remaining atomic questions and enforcement under failure have not been established by a bounded workflow test.

Source: https://docs.replit.com/features/collaboration/team-workspaces
Locator: Personal vs. team workspaces; Managing members
Finding: Personal workspaces have an admin and project guests; team workspaces add members with access to all workspace projects. Admins can invite and remove members.
Retrieved: 2026-11-08

Source: https://docs.replit.com/teams/configure-saml-sso
Locator: Configure single sign-on
Finding: SAML SSO requires Enterprise and organization/identity-provider admin access. SSO does not automatically invite users; invitations or SCIM manage membership.
Retrieved: 2026-11-08

## replit:TEAM-03

Finding: Humans approve plans before background execution and apply or dismiss finished changes before they affect the main version.

Verification: agent-cross-checked
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation evidence only. Remaining atomic questions and enforcement under failure have not been established by a bounded workflow test.

Source: https://docs.replit.com/core-concepts/agent/task-system
Locator: Review and accept; Review and apply
Finding: Humans approve plans before background execution and apply or dismiss finished changes before they affect the main version.
Retrieved: 2026-11-08

## replit:TEAM-05

Finding: Replit displays code diffs in its Changes pane, alongside local changes or linked PRs, in addition to task logs, tests and previews for review decisions.

Verification: sourced
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation and pinned source inspection only; no hands-on scenario was run.
Other atomic questions remain unestablished: evidence_navigation, comments, request_changes, effort, missed_evidence, keyboard, narrow_screen, context_switching, roles, retained_context.

Source: https://docs.replit.com/core-concepts/agent/task-system
Locator: Review and apply
Finding: Ready tasks expose a work log, test results and live preview with Apply or Dismiss decisions. Review effort and missed evidence were not measured.
Retrieved: 2026-11-08

Source: https://docs.replit.com/build/code/navigate-your-workspace.md
Locator: Changes and pull requests
Finding: The product Changes pane switches between local changes and linked pull requests, shows branch/description details and renders the proposed code diff with added/removed lines. Linked requests also appear in Details and review feedback can be addressed through Agent.
Retrieved: 2026-11-13

## replit:SHIP-01

Finding: Task review retains a work log, test results and a live preview in the task flow. A cross-task artifact gallery was not established.

Verification: agent-cross-checked
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation evidence only. Remaining atomic questions and enforcement under failure have not been established by a bounded workflow test.

Source: https://docs.replit.com/core-concepts/agent/task-system
Locator: Review and apply
Finding: Task review retains a work log, test results and a live preview in the task flow. A cross-task artifact gallery was not established.
Retrieved: 2026-11-08

## replit:SHIP-02

Finding: With App Testing enabled, Agent periodically tests apps it builds in an actual browser and fixes issues it finds. It is available for Full Stack JavaScript and Streamlit Python web apps. Documentation is not proof of mandatory independent review or an enforced passing test gate.

Verification: sourced
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

App Testing covers Full Stack JavaScript and Streamlit Python web apps only, and Agent decides when to test rather than testing after every message.
Documentation evidence only. Remaining atomic questions and enforcement under failure have not been established by a bounded workflow test.

Source: https://docs.replit.com/features/agent/app-testing
Locator: App Testing
Finding: With App Testing enabled, Agent periodically tests apps it builds in an actual browser and fixes issues it finds; it is available for Full Stack JavaScript and Streamlit Python web apps. Documentation is not proof of mandatory independent review or an enforced passing test gate.
Retrieved: 2026-11-08

## replit:SHIP-03

Finding: The repository contribution workflow creates branches, commits/pushes code and opens draft GitHub PRs or GitLab MRs, then can revise the same request after review.

Verification: sourced
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation and pinned source inspection only; no hands-on scenario was run.
Other atomic questions remain unestablished: worktrees, ci_response, rebase, product_state.

Source: https://docs.replit.com/learn/build-in-parallel
Locator: Apply changes safely
Finding: Applying tasks invokes AI-assisted conflict resolution. The source does not establish a GitHub PR workflow.
Retrieved: 2026-11-08

Source: https://docs.replit.com/build/code/working-with-code.md
Locator: Sandbox architecture; draft PR and review workflow
Finding: The current repository-contribution guide explicitly says the development sandbox runs in a microVM. Replit owns the contribution workflow: it creates a branch, commits/pushes changes, opens a GitHub draft PR or GitLab MR, and updates the request after review.
Retrieved: 2026-11-13

Source: https://docs.replit.com/build/code/navigate-your-workspace.md
Locator: Changes and pull requests
Finding: The product Changes pane switches between local changes and linked pull requests, shows branch/description details and renders the proposed code diff with added/removed lines. Linked requests also appear in Details and review feedback can be addressed through Agent.
Retrieved: 2026-11-13

## replit:SHIP-04

Finding: Users publish to a live URL from the Project Editor. Production authorization and revision binding were not tested.

Verification: agent-cross-checked
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation evidence only. Remaining atomic questions and enforcement under failure have not been established by a bounded workflow test.

Source: https://docs.replit.com/learn/projects-and-artifacts/project-editor
Locator: How do I publish my app?
Finding: Users publish to a live URL from the Project Editor. Production authorization and revision binding were not tested.
Retrieved: 2026-11-08

## replit:SHIP-05

Finding: Agent creates checkpoints that users can roll back to when a change is wrong.

Verification: agent-cross-checked
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation evidence only. Remaining atomic questions and enforcement under failure have not been established by a bounded workflow test.

Source: https://docs.replit.com/features/agent/overview
Locator: What happens if Agent makes a mistake?
Finding: Agent creates checkpoints that users can roll back to when a change is wrong.
Retrieved: 2026-11-08

## replit:OWN-01

Finding: The main Replit Agent/platform is proprietary software supplied under Replit service terms, which reserve its intellectual-property rights and grant users limited service access.

Verification: sourced
Checked: 2026-11-13
Scheduled recheck: 2027-05-12
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation and pinned source inspection only; no hands-on scenario was run.
Other atomic questions remain unestablished: client_license, hosted_extensions, source_revision, hosted_parity.

Source: https://replit.com/terms-of-service
Locator: Replit Ownership; limited access rights
Finding: The current service terms reserve all rights in the Replit Service and related software to Replit and give users only the expressly granted limited access rights. This classifies the main Agent/platform as proprietary, separately from user-created code and any separately licensed public client/SDK components.
Retrieved: 2026-11-13

## replit:OWN-02

Finding: Replit imports/exports project artifacts and offers Enterprise deployment options, but even Full BYOC explicitly leaves the control plane external and does not provide an entirely self-hosted Replit installation.

Verification: sourced
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation and pinned source inspection only; no hands-on scenario was run.
Other atomic questions remain unestablished: export, formats, cancellation, account_independence, history_portability.

Source: https://docs.replit.com/learn/projects-and-artifacts/project-editor
Locator: Import code and design
Finding: Users can import GitHub projects and uploaded files. Exported task history and complete self-hosting are not established.
Retrieved: 2026-11-08

Source: https://docs.replit.com/teams/bring-your-own-cloud.md
Locator: Compare deployment options; Full BYOC warning and availability
Finding: Full BYOC includes Agent/development execution in a customer tenant but is explicitly private preview, with eligibility/configurations confirmed through Enterprise onboarding. Published-app BYOC and Image Exports are early access and do not move the development/Agent control plane. Even Full BYOC leaves the control plane external and explicitly is not an entirely self-hosted or air-gapped Replit installation.
Retrieved: 2026-11-13

## replit:OWN-03

Finding: Pro users can opt out of Model Improvement in Personalization settings, and Replit states it does not train on Enterprise customer content. No such control is named for Starter or Core. This is distinct from telemetry opt-out.

Verification: sourced
Checked: 2026-11-08
Scheduled recheck: 2026-12-08
Plan: Pro opt-out; Enterprise excluded from training
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation evidence only. Remaining atomic questions and enforcement under failure have not been established by a bounded workflow test.

Source: https://docs.replit.com/legal-and-security-info/model-improvement
Locator: Model improvement
Finding: Pro users can opt out of Model Improvement in Personalization settings; Replit states it does not use Enterprise customer content to train models. The page names no such control for Starter or Core.
Retrieved: 2026-11-08

## replit:OWN-04

Finding: Replit supplies recurring Core/Pro plans and metered Agent/cloud credits. The default public Core plan is USD 20 per month at monthly billing; regional offers can differ. Pro has no per-user fees. App BYOK is separate from the managed Agent model and no one-time software license is offered.

Verification: sourced
Checked: 2026-11-08
Scheduled recheck: 2026-11-22
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Current public pricing page and its immutable first-party pricing catalog retrieved 2026-11-13

Documentation and pinned source inspection only; no hands-on scenario was run.
Other atomic questions remain unestablished: seat_minimum, workspace_fees, included_usage, trial, taxes.

Source: https://docs.replit.com/billing/ai-billing
Locator: Getting started; Managing your spend
Finding: Monthly plan credits cover Agent and other cloud services such as published apps, storage and databases under usage-based billing, with alerts, budget limits and purchasable credit packs. The page does not list network transfer or state an overage rate.
Retrieved: 2026-11-08

Source: https://docs.replit.com/billing/plans/starter-plan
Locator: Upgrading to Core
Finding: Upgrading to Core gives more storage and monthly credits that also let users pay as they go.
Retrieved: 2026-11-08

Source: https://docs.replit.com/help/agent-and-ai.md
Locator: Can I use my own API keys with Agent?
Finding: The first-party FAQ separates app API credentials from the Agent itself: user provider keys in Secrets power calls made by the built app, while the Agent experience uses Replit-managed model infrastructure rather than the user-key route. Curated in-product model selection is separate from replacing that infrastructure with an independently keyed endpoint. Replit AI billing charges its own plan credits. This is not evidence of a third-party coding harness or a native subscription-login path.
Retrieved: 2026-11-13

Source: https://docs.replit.com/billing/plans/replit-pro.md
Locator: Tiered monthly credits; up to 15 builders; no per-user fees
Finding: The current Pro documentation explicitly supplies a flat recurring team plan with tiered monthly credits and up to 15 builders, with no per-user fees. Core/Pro subscriptions and usage credits are documented; one-off credit-pack top-ups are prepaid consumption rather than a one-time software license.
Retrieved: 2026-11-13

Source: https://replit.com/pricing
Locator: Monthly/Yearly Core, Pro and Enterprise plan offering
Finding: The current public plan offering lists recurring Core and Pro subscriptions and an Enterprise sales plan. It does not offer the whole ADE free without a paid tier or a one-time software license. The page defaults to annual-billing display; this pass does not convert annual prices or infer a USD monthly price from a dollar sign.
Retrieved: 2026-11-13

Source: https://cdn.replit.com/_next/static/chunks/2kybev8fpsqw3.js
Locator: Pricing catalog module 596139; Core default monthly Stripe price and USD allocation
Finding: The immutable first-party script loaded by the public pricing page defines the non-legacy default Core monthly price as 2000 cents with a 20-USD monthly allocation. The old 2500-cent option is explicitly legacy-only. Annual pricing and restricted regional Core offers are separate entries. This establishes the default public Core plan at USD 20 per month without converting an annual rate.
Retrieved: 2026-11-13

## replit:OWN-06

Finding: The November 6 changelog documents Agent-assisted support investigation and Pro ticket submission. This is shipped release activity, not a shutdown notice.

Verification: agent-cross-checked
Checked: 2026-11-08
Scheduled recheck: 2026-11-22
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation evidence only. Remaining atomic questions and enforcement under failure have not been established by a bounded workflow test.

Source: https://docs.replit.com/updates/2026/11/06/changelog
Locator: Get support directly from Agent
Finding: The November 6 changelog documents Agent-assisted support investigation and Pro ticket submission. This is shipped release activity, not a shutdown notice.
Retrieved: 2026-11-08

## replit:OWN-08

Finding: Starter offers recurring daily Agent credits capped monthly and monthly cloud credits, with one published app whose public link expires after thirty days. Full build, connectors and Plan mode require Core.

Verification: agent-cross-checked
Checked: 2026-11-08
Scheduled recheck: 2026-11-22
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation evidence only. Remaining atomic questions and enforcement under failure have not been established by a bounded workflow test.

Source: https://docs.replit.com/billing/plans/starter-plan
Locator: What you get; Limitations
Finding: Starter offers recurring daily Agent credits capped monthly and monthly cloud credits, with one published app whose public link expires after thirty days. Full build, connectors and Plan mode require Core.
Retrieved: 2026-11-08

## replit:OWN-09

Finding: Agent uses effort-based paid pricing with plan credits; actual charges depend on request complexity and cloud services. A fixed average task price is not documented.

Verification: agent-cross-checked
Checked: 2026-11-08
Scheduled recheck: 2026-11-22
Plan: Not specified
Surface: Replit hosted Project Editor and Agent; mobile apps and paid team features where named
Version: Public documentation retrieved 2026-11-08

Documentation evidence only. Remaining atomic questions and enforcement under failure have not been established by a bounded workflow test.

Source: https://docs.replit.com/billing/ai-billing
Locator: Effort-based pricing
Finding: Agent uses effort-based paid pricing with plan credits; actual charges depend on request complexity and cloud services. A fixed average task price is not documented.
Retrieved: 2026-11-08