Ona is the new name that Gitpod, long known for cloud development environments, adopted in October 2025. Instead of running on your own machine, its AI agents run in the cloud: hand over a task and the agent works autonomously from code changes through to opening a pull request. What sets it apart is not only how fast agents run, but the controls that let a company hand work to them with confidence. Dedicated cloud environments, network controls, audit trails, and SOC 2 / GDPR readiness are built into the foundation. In June 2026 OpenAI announced its acquisition of Ona (subject to regulatory approval), positioning it as infrastructure behind Codex.
Key Features
- Background agents: The stated flow is “task in, pull request out.” Because processing runs in the cloud, work continues even after you close your laptop. Documented use cases include CVE remediation, code review, summarizing CI failures, legacy code migration, and bug triage
- Automations for event-driven runs: Agents can be triggered by pull requests, schedules, webhooks, or issue tracker events. This makes it practical to roll the same change across multiple repositories as a single body of work
- Reproducible development environments: Environments are defined as code on top of Dev Containers. This strength carries over from Gitpod and avoids the “works on my machine” problem caused by environment drift. Higher configurations offer up to 32 cores, 128GB RAM, and 200GB of disk, with GPU support
- Runtime AI security: Deployment into your own VPC, network control, scoped credentials, kernel-level policy enforcement, and detailed audit trails. The design lets you trace after the fact what an agent accessed
- Integration with existing tools: Beyond source control such as GitHub and GitLab, it connects to editors including Claude Code and Cursor, plus tools like Jira and Notion. It also supports the Model Context Protocol (MCP), with organization-side control over how MCP is used
- Cloud or your own environment: Ona Cloud offers the quickest start, and self-hosting inside your own VPC on AWS or GCP is also supported
Pricing
| Plan | Price | What’s included |
|---|---|---|
| Core | From $20/month | 80–2,200 OCUs (Ona Compute Units) per month, up to 100 seats, unlimited parallel environments, prebuilds, role-based access management |
| Add-on credits | $10 / 40 OCUs | Additional credits for the Core plan |
| Enterprise | Contact sales | Everything in Core plus VPC deployment, network control, SSO and OIDC, organization-wide secrets, audit trails, warm pools, MCP usage control, API/SDK access, SLA, and a dedicated account manager |
The official pricing page lists no free plan. Usage is metered in credits called OCUs, and on the Core plan environments are automatically deleted after seven days of inactivity.
Pricing is current as of August 2026. Please check the official site for the latest information.
Pros & Cons
✅ Pros
- Agent work completes in the cloud, so long-running tasks are not tied to your local machine
- Governance features that corporate IT teams look for — audit trails, permission controls, VPC deployment — are available from the start
- Built on the proven foundation of environment definitions via Dev Containers, carried over from the Gitpod era
- Automations can handle routine work, making it suited to ongoing operation rather than one-off requests
- Connects to GitHub, GitLab, Jira, and existing AI coding tools, so it can be adopted without overhauling your development flow
⚠️ Cons
- No free plan, so it is not well suited to individuals who simply want to try it out
- Because pricing is credit-based via OCUs, the actual monthly cost is hard to predict from usage alone
- Core governance features such as network control and auditing sit mainly on the Enterprise plan
- After the OpenAI acquisition closes, the delivery model and its standing as a standalone service may change
- The design assumes work is delegated to agents, so results depend on having review practices and tests in place
Comparison with Similar Services
| Criteria | Ona | Devin | GitHub Copilot (coding agent) | Cursor |
|---|---|---|---|---|
| Provider | Gitpod GmbH (acquisition by OpenAI announced) | Cognition | GitHub | Anysphere |
| Where it runs | Cloud (self-hosting in your own VPC also possible) | Cloud | GitHub’s cloud environment | Mainly a local editor |
| Main positioning | Execution platform and governance for agents | Autonomous AI software engineer | Agent tied to a repository | AI-powered code editor |
| Governance and auditing | VPC, audit trails, and policy enforcement built in | Enterprise features available | Follows GitHub’s permission model | Limited organization-level features |
| Free plan | None | None | Included in paid plans | Available |
Who Is It For
- Teams that want to process routine changes spanning multiple repositories at once, such as CVE remediation or library migrations
- Companies for which permission, audit, and network controls are internal requirements for adopting AI agents
- Organizations already using Gitpod for cloud development environments that want to extend into agent operations
- Development teams that need long-running background processing that continues after the laptop is closed
- Environments with strict security requirements that need agents to run inside their own VPC
Summary
Ona is a platform that combines the execution environment and the governance needed to delegate development work to AI agents. Built on the cloud development environment technology accumulated during the Gitpod years, it pairs agents running in the background with auditing and permission management. With no free plan and a design aimed at enterprise use, it fits teams running repeated maintenance work more than individuals experimenting on their own. The effects of the OpenAI acquisition announced in June 2026 will become clearer over time, so check the official site for the latest availability before committing.