Replit Agent / Code Execution API

license

Proprietary

source

closed

service

proprietary(Replit managed Autoscale Deployments, no self-host)

open-part

third-party isolation primitive omegajail(github.com/omegaup/omegajail) + an ARCHIVED thin client (replit-code-exec, ISC, read-only since 2025-02-11)

Replit's only published client, replit-code-exec, is archived and read-only, and its README describes forking a Repl and deploying it onto Replit's own Autoscale Deployments rather than running a self-hostable service. The container sandbox behind it, omegajail, is open source, but it is a third-party dependency rather than anything Replit publishes as its own open core.

• https://blog.replit.com/ai-agents-code-execution recorded 2026-06-04

  canonical Replit blog: Code Execution API uses omegajail on Autoscale Deployments, ~100ms, Python-focused (Feb 2024)

• https://github.com/replit/replit-code-exec recorded 2026-08-13

  Replit code-exec client: stateless API for AI agents, runs untrusted Python in ephemeral unprivileged containers using omegajail (github.com/omegaup/omegajail); ISC license; repo ARCHIVED read-only 2025-02-11

• https://pypi.org/pypi/replit-code-exec/json recorded 2026-09-29

  PyPI metadata for the client package: the description says the API executes code "inside an ephemeral" container and "created on the fly running in Replit Deployments using", with "https://github.com/omegaup/omegajail as code sandbox", the license field is "ISC", and setting up a copy means forking a Repl ("Fork it to your" account) then deploying it, "This is only compatible with the Autoscale Deployments."

• https://raw.githubusercontent.com/replit/replit-code-exec/main/LICENSE recorded 2026-09-29

  The client repository LICENSE is the ISC text, "Copyright 2023 Replit" with "Permission to use, copy, modify, and/or distribute this software for any purpose", covering only the thin client library.

• https://raw.githubusercontent.com/replit/replit-code-exec/main/README.md recorded 2026-09-29

  The client README says the package is "An interface to interact with Replit" and that the only way to stand up a server is to open the eval-python Repl and "Fork it to your" account, then deploy it, noting "This is only compatible with the Autoscale Deployments.", so no off-platform, self-hostable service exists; searched for self-host, on-prem, Docker Compose, enterprise and license key: none.

• https://replit.com/blog/ai-agents-code-execution recorded 2026-06-04

  Code Execution API uses omegajail (open source unprivileged container sandbox) on Replit's proprietary Autoscale Deployments; replit-code-exec PyPI client; Python-focused

• https://replit.com/terms-of-service recorded 2026-09-29

  Replit Terms of Service ownership clause: "Replit owns and retains all right, title, and interest in and to the Service and related software", and "you obtain no rights to the Service or intellectual property rights by implication, estoppel, or otherwise", while the acceptable-use list bars "reverse-engineering any part of the Service", so the hosted execution service is proprietary and closed.

Replit discloses no standalone usage figure for the code-execution API, so this reading follows the wider Replit Agent and platform surface it powers, which runs many live projects on Google Cloud Run. Neither Replit nor the API publishes a countable artifact of its own, so the level follows those platform figures rather than a verifiable number, and the API alone would likely place lower.

• https://cloud.google.com/blog/products/serverless/whats-new-for-cloud-run-at-next26 recorded 2026-08-13

  Replit runs >1M live projects (on Cloud Run), primary-source scale signal for the Replit surface

• https://replit.com/products/agent recorded 2026-06-04

  Replit Agent product page (the agent surface the code-exec API backs)

Replit runs each execution inside an ephemeral, unprivileged container on its own Deployments infrastructure, sandboxed by the third-party omegajail project, as a stateless API purpose-built for AI agents with a swappable image. Isolation stays at the container level rather than a microVM, the language focus is narrow (Python), and sandboxes carry no persistence between calls.

• https://blog.replit.com/ai-agents-code-execution recorded 2026-09-29

  Canonical blog host redirects to the replit.com post and serves the same page, including "Requests are handled in as little as 100ms and use the" omegajail sandbox.

• https://pypi.org/pypi/replit-code-exec/json recorded 2026-09-29

  PyPI metadata for the client package: the description says the API executes code "inside an ephemeral" container and "created on the fly running in Replit Deployments using", with "https://github.com/omegaup/omegajail as code sandbox", the license field is "ISC", and setting up a copy means forking a Repl ("Fork it to your" account) then deploying it, "This is only compatible with the Autoscale Deployments."

• https://raw.githubusercontent.com/replit/replit-code-exec/main/README.md recorded 2026-08-14

  stateless API optimized for AI agents, executing untrusted Python inside an ephemeral unprivileged container created on the fly on Replit Deployments using omegaup/omegajail as the code sandbox; the container image is swappable via `evalctl image`; Autoscale Deployments only; no latency figure stated.

• https://replit.com/blog/ai-agents-code-execution recorded 2026-09-29

  Replit blog post on the Code Execution API: "Requests are handled in as little as 100ms and use the" omegajail sandbox, the API is "a stateless API container server that is deployable through" Replit Autoscale Deployments, the image is customizable ("You can easily customize the Docker container image"), it is Python-focused ("This solution works best for simple math evaluation using Python"), and the stateful variant is discontinued ("We've concluded this experiment and have decided not to support it at this time.").