AI CodeOps

About AI CodeOps

AI CodeOps (formerly CodeOps Studio) is a browser-based platform where AI agents plan, build, review, test and ship code, and nothing merges until quality gates and a human approve it. It is built and operated by TechInnoHub.

Our name

AI CodeOps was called CodeOps Studio until October 6, 2026, when it moved from codeopsstudio.dev to aicodeops.dev (the announcement). Old links to codeopsstudio.dev still work: each one redirects to the same page here. AI CodeOps is not affiliated with the open-source CodeOps Studio mobile IDE for Android, or with other products that use the CodeOps name.

Why we're building it

AI can now write a working change in minutes. The hard part moved: knowing whether that change is right, whether it was reviewed, whether the tests really ran, what it cost, and who approved it. Most teams answer those questions by hand, after the fact, if at all.

Matt started AI CodeOps to put those answers into the workflow itself. A session in AI CodeOps doesn't end when the code is written: specialised agents review it, the project's real tests run, the release gate decides whether it can merge, and the whole history is kept as a signed evidence pack anyone can verify.

How a change moves through AI CodeOps

  1. You describe a task and pick a workflow and a project. Before anything runs, AI CodeOps checks the workspace budget and policy: a deny rule stops the session, and an ask rule holds it for approval.
  2. Agents work through the workflow. The default workflow runs four steps in order: the Architect plans, the Builder builds, the Reviewer reviews and the Tester tests. Each step sees the task and what the earlier steps wrote.
  3. The change is written to a fresh copy of your repository, producing a diff.
  4. Your tests run in an isolated sandbox that is cut off from the network, using the project's test command. The result is recorded: passed, failed, skipped and coverage.
  5. The diff is reviewed. A review pass reads the actual diff and records findings, each with a severity. If the review can't finish, the change is held, because its findings are unknown.
  6. A person approves. The person who started the session can't approve it while someone else in the workspace can.
  7. The pull request opens only when the policy, the approval and the quality gate all pass: no open high or critical findings, no failing tests and a completed review. If you merge from the console, the gate is checked again first. After the merge, a project with a deploy target has its deploy webhook called and the deployment recorded.

The how it works guide follows a session through each stage in more detail.

How we work

  • In the browser. Nothing to install. Sessions run on our servers; tests and terminals run in isolated sandboxes.
  • Your Git host. GitHub, GitLab, Bitbucket, Gitea or Forgejo, including self-hosted instances. See integrations.
  • Honest about status. AI CodeOps is in open beta. We publish what works on the changelog and live health on the status page, and we don't claim certifications we don't have.
  • Costs you can see. Every session shows what it cost, and workspaces can cap their spend. See pricing.
  • Your data isn't training data. Our data processing agreement says we process workspace content only to provide the service, and that we don't use it to train AI models.

What AI CodeOps doesn't do

Knowing the limits saves you time, so here they are:

  • There is no IDE extension, CLI or desktop app. Everything happens in the browser.
  • It doesn't connect to Azure DevOps.
  • It reviews the changes made in its own sessions, not pull requests people open directly on the Git host.
  • It doesn't deploy your code itself. It calls a deploy webhook you set and records the deployment; your own pipeline does the deploy.
  • It doesn't generate SBOMs or scan dependencies for known vulnerabilities. You record the SBOMs your own tools produce, and AI CodeOps uses them for license checks.

Plans

AI CodeOps is free during the open beta, with a monthly AI allowance on the platform's models; a workspace can also bring its own provider key. Enterprise is priced per seat through sales and adds single sign-on with SAML 2.0 or OpenID Connect, custom roles, audit history kept for as long as your organization exists, and self-hosted deployment on request, which is available only on Enterprise. See pricing and Enterprise.

The company

AI CodeOps is operated by TechInnoHub. The terms of service are a contract between you and TechInnoHub. Under the privacy policy, TechInnoHub is the controller of the personal data in your account; for the content of a workspace that belongs to an organization, it acts as that organization's processor under the data processing agreement. Logos, screenshots and a short company description are in the press kit.

Get in touch

Questions, feedback or partnership ideas: see contact. For security, read how we protect your code.

Product support and security reports go to [email protected] (put "Security" in the subject for a vulnerability report). Signed in? Use Help in any workspace. For Enterprise plans, SSO or self-hosted deployment, use the Enterprise contact form.

Follow us on LinkedIn and Medium.