# OpenComputer vs Ellipsis

Ellipsis (ellipsis.dev) and OpenComputer both run Claude Code and Codex in the cloud, take requests from Slack, and open pull requests. Neither charges per seat.

The difference is what you are building. Ellipsis Agent Cloud is a platform for defining agents as code: each agent is a YAML file in your repository with its model, prompt, triggers, permissions and budget, deployed when you merge, and billed on usage from prepaid credits. OpenComputer is a ready-made coding agent you mention in Slack, built around a shared thread the whole team can steer, and billed per workspace for the minutes it works.

If you want to configure a fleet of governed agents triggered by GitHub, Linear, Sentry and cron, with spend caps on each, Ellipsis fits. If you want a teammate in Slack that turns requests into pull requests without writing configuration first, OpenComputer fits.

## At a glance

| | Ellipsis | OpenComputer |
|---|---|---|
| Where the agent runs | Isolated cloud sandbox per session | Your workspace's cloud Linux machine |
| Agent and models | Claude Code and Codex, models you set per agent | Claude Code and Codex |
| How you hand it work | Slack, GitHub, Linear, Sentry, cron, CLI, API, SDKs, MCP | Slack mention or API |
| What you get back | PRs, code reviews and session transcripts | A pull request with screenshot and preview for UI changes |
| Between tasks | Cached base build; paused environments not billed | Machine stays warm; repo and caches kept |
| Pricing model | Usage-based credits: tokens, compute, platform fee | Per workspace, work minutes |
| Entry price | No plan fee; pay for usage | Pro, $20 per month for the workspace |
| Free option | Case-by-case free usage for open source | 3-day trial, 600 work minutes |

## What is Ellipsis?

Ellipsis began in 2023 as an AI code-review bot. In July 2026 it [introduced the Ellipsis Agent Cloud](https://www.ellipsis.dev/blog/the-ellipsis-agent-cloud), a managed platform that runs coding agents in the cloud, along with an agent CLI and usage-based pricing with no seats.

[Every agent session runs in its own sandbox](https://www.ellipsis.dev/pricing). An [environment file](https://www.ellipsis.dev/docs/environments.md) in `.ellipsis/environments/` sets which repositories are checked out, scripts, variables, secrets and MCP servers. Its `build_base` hook installs dependencies and is cached until the input files you list change. The [environment schema](https://www.ellipsis.dev/docs/environments/schema.md) allows 2 to 32 vCPUs, up to 64 GB of memory, and a session timeout of up to one hour.

Agents are YAML files that sync from your default branch, so changing an agent is a pull request. Sessions start from [`@ellipsis` mentions in Slack, GitHub or Linear, Sentry alerts, cron schedules and GitHub events](https://www.ellipsis.dev/docs/triggers.md), or from the dashboard, the CLI, the REST API, the Python and TypeScript SDKs, or an MCP server. Each session gets a [scoped GitHub token that is revoked when it ends](https://www.ellipsis.dev/docs/sessions/lifecycle.md), and [budgets](https://www.ellipsis.dev/docs/budgets.md) cap spend per session, per agent and per account. Code review is still built in, with one review per push written by reviewers you define.

## What is OpenComputer?

OpenComputer runs Claude Code and Codex in its own cloud, and you hand them work by mentioning @Computer in a Slack thread or calling the API from your code. Each Slack workspace gets its own cloud Linux machine, sized to its plan. Every thread works in its own copy of the repos on its own branch, so parallel requests don't collide. Workspaces resume warm, with the repository, dependencies and build caches already in place.

@Computer reads the code, runs the commands and tests, then opens a pull request that says what changed, why, and how it was checked, with a link back to the thread. UI changes come with a screenshot and a preview link. For teams that want their own agents, OpenComputer also runs [serverless agents](/agents) written in TypeScript, on the same work minutes.

## Key differences

### Configure agents or mention one

Ellipsis asks you to define the work up front. You write an environment file and agent files, review them in a PR, and from then on agents fire on the triggers you set. That suits recurring jobs such as triaging Sentry alerts, and keeps every agent's behavior versioned in git.

OpenComputer starts from a request. Someone mentions @Computer in a thread, and it reads the code, does the work and opens a pull request. There is no agent file to write. When you do want a custom agent on a schedule or webhook, OpenComputer's serverless agents are TypeScript functions rather than YAML, with durable sessions, memory across sessions and Slack, GitHub and Linear connections.

### Governance and controls

Ellipsis puts governance at the center. Per-session, per-agent and per-account budgets are enforced, permissions are scoped per agent, GitHub tokens are minted per session, and every session has a full transcript. The default and maximum budget per session is $250.

OpenComputer's controls are simpler. Work minutes are capped by your plan and prepaid credits, and every file @Computer opens and every command it runs appears in the Slack thread. For serverless agents, API keys and secrets never enter the agent's runtime: the platform holds them and injects them on outbound requests.

### Team workflow in Slack

Both take mentions in Slack. On OpenComputer the thread is the unit of work: everyone in it sees what @Computer is doing, anyone can steer it, and conventions one person explains apply to everyone's later requests. The thread is also the record of who asked for each change and why. Ellipsis treats Slack as one trigger among many, alongside GitHub, Linear and Sentry, and you can send follow-up messages to a session mid-task.

### Models and billing

On Ellipsis, models are set per agent, and [Claude Code agents](https://www.ellipsis.dev/docs/harnesses/claude.md) support effort levels and fallback models. Usage is paid from [prepaid credits that do not expire](https://www.ellipsis.dev/docs/pricing.md): model tokens at the provider's rates, sandbox compute, and a platform fee. You can also deploy into your own AWS account, where AWS bills tokens and compute. OpenComputer lets you bring your own Claude or ChatGPT subscription or API key for $0 in model costs, or use its tokens at API prices with no markup.

## Pricing compared

As of October 2026, Ellipsis has [no tiers, feature gates or seat fees](https://www.ellipsis.dev/pricing). You pay for usage from prepaid credits, covering model tokens, compute and a platform fee, with optional auto-recharge. Paid support packages start at $5,000 a month, and open source projects can ask for free usage case by case. OpenComputer lists Pro at $20 a month with 2,000 work minutes and Max at $200 with 25,000, for the whole Slack workspace, and $0.01 per extra minute.

A side-by-side example would not be meaningful: an Ellipsis bill depends on the tokens each session uses, the sandbox size and how long it runs, while an OpenComputer bill depends only on work minutes when you bring your own model subscription. Check the current rates on [Ellipsis's pricing page](https://www.ellipsis.dev/pricing) against your expected usage.

## When to choose Ellipsis

- You want agents defined as code, versioned in your repo and changed through pull requests.
- You need hard spend caps per session, per agent and per account, plus scoped permissions and transcripts.
- You want every agent session in its own sandbox.
- You want agents triggered by Sentry alerts, GitHub events, Linear and cron without writing code.
- You want code review on every push from the same platform, or a deployment in your own AWS account.

## When to choose OpenComputer

- Your team works in Slack and wants requests, steering and history in one shared thread.
- You want to start handing work to an agent without writing configuration first.
- You want to use your Claude or ChatGPT subscription and pay a predictable plan per workspace.
- You want to build custom agents in TypeScript, such as [a background coding agent](/blog/background-coding-agent), on the same minutes.

## FAQ

### Is OpenComputer an Ellipsis alternative?

Yes, for teams that want Claude Code or Codex to take coding requests from Slack and open pull requests. Ellipsis is a platform for configuring many governed agents as code, while OpenComputer is a ready-made Slack agent with per-workspace pricing. Teams comparing agent platforms may also want to read [OpenComputer vs Tembo](/compare/opencomputer-vs-tembo).

### Is Ellipsis still an AI code review tool?

Code review is still part of Ellipsis, with one review per push written by reviewers you define. Since July 2026 the main product is the Ellipsis Agent Cloud, which runs coding agents in cloud sandboxes. Many third-party comparisons still describe only the older code-review product and its seat pricing.

### How much does Ellipsis cost?

Ellipsis has no seat fees and no plan tiers. You buy prepaid credits that pay for model tokens at provider rates, sandbox compute and a platform fee. Current compute rates and fee details are on its pricing page.

### Can I use my Claude or ChatGPT subscription with OpenComputer?

Yes. OpenComputer runs Claude Code and Codex on your Claude or ChatGPT subscription for $0 in model costs, or on your own API key. When a subscription hits its limit, you choose whether to switch to OpenComputer tokens at API prices or wait for the reset.

### How long can an Ellipsis session run?

The Ellipsis environment schema sets a sandbox timeout between 60 seconds and one hour, with one hour as the default. Sessions move through queued, preparing, running, idle and closed states, and queue when your account reaches its live-session limit.

## Sources

- [Ellipsis homepage](https://www.ellipsis.dev)
- [Introducing the Ellipsis Agent Cloud](https://www.ellipsis.dev/blog/the-ellipsis-agent-cloud)
- [Ellipsis pricing](https://www.ellipsis.dev/pricing)
- [Pricing docs](https://www.ellipsis.dev/docs/pricing.md)
- [Environments](https://www.ellipsis.dev/docs/environments.md)
- [Environment schema](https://www.ellipsis.dev/docs/environments/schema.md)
- [Triggers](https://www.ellipsis.dev/docs/triggers.md)
- [Session lifecycle](https://www.ellipsis.dev/docs/sessions/lifecycle.md)
- [Budgets](https://www.ellipsis.dev/docs/budgets.md)
- [Claude Code harness](https://www.ellipsis.dev/docs/harnesses/claude.md)