Description
Ellipsis is a cloud platform for running coding agents built on Claude Code or Codex. Instead of installing an agent locally, teams configure it once, in a YAML file kept in their own repository, and Ellipsis runs it in an isolated sandbox whenever it is triggered.
Read More
Setup happens in three stages. Teams define an environment, specifying which repository to use, its dependencies, environment variables and compute resources. Agents then spawn in an isolated cloud sandbox, invoked through the API, a schedule, or an event such as a GitHub comment, a Slack message or a Linear ticket. Every session is logged: commands, tool calls and results are recorded, along with token cost, so a team can review exactly what an agent did afterward.
Because agents run in short-lived sandboxes, Ellipsis scopes a fresh set of credentials to each session and revokes them when the session ends, and it states it retains no source code once a session closes, keeping only the audit trail. Agents can be reached with @ellipsis mentions on GitHub, Slack and Linear and answered inline; a single agent stays attached to a PR, issue or thread as a durable conversation across multiple exchanges. A shared skill library lets multiple agents reuse the same instructions across repositories, and a template gallery can deploy a working agent in one command. A REST API and CLI expose everything the dashboard can do, and a running session can be steered live with a new message mid-task.
Ellipsis is free for individuals who already have a Claude Code or Codex subscription. For organizations, pricing is usage-based: token cost plus a 10% platform fee, plus compute billed by vCPU-hour and GB-hour, with no per-seat fees. Teams can run on Ellipsis's managed cloud or deploy the same agents into their own AWS account, where the control plane and sandboxes stay inside their VPC. Optional paid support packages, starting at $5,000 a month, add a dedicated Slack channel, faster response times and help wiring up custom agents.
It is built for software teams that already use Claude Code or Codex and want those agents running unattended in the cloud, triggered by their existing GitHub, Slack, Linear and Sentry workflows rather than a developer's own machine. Ellipsis states it is SOC 2 Type I certified, with Type II in progress.
- CategoryAI Code Review & Testing
- Pricing ModelFreemium
- DeploymentCloud
- PlatformsWeb
Ellipsis's Core Features
YAML-defined agents stored and versioned in the team's own repository
Isolated cloud sandbox per agent session
Triggers via API calls, cron schedules, or GitHub, Linear, Slack and Sentry events
@ellipsis mentions answered inline on GitHub, Slack and Linear
Durable, persistent conversations tied to a single PR, issue or thread
Scoped, per-session credentials revoked when the sandbox closes
Full audit trail of commands, tool calls and cost after each session
Shared skill library across agents and repositories
REST API and CLI for scripting anything the dashboard does
Option to deploy into a team's own AWS account
How to use Ellipsis?
Connect a repository: Install Ellipsis on GitHub and pick the repositories agents should work in.
Define an agent: Write a YAML config specifying the environment, dependencies and triggers.
Choose a trigger: Set the agent to run on a schedule, a GitHub, Linear or Slack event, or an @ellipsis mention.
Let the agent run: Ellipsis spins up an isolated sandbox, and the agent opens PRs, answers questions or investigates issues.
Review the session: Check the transcript, tool calls and token cost for each run in the dashboard.
Scale to your own cloud: Optionally move the same agent configs to run inside your own AWS account.
Ellipsis's Use Cases
- Automated PR review
- Scheduled maintenance agents
- In-thread codebase Q&A
- Production issue investigation
