Ellipsis AI Code Reviews & Bug Fixes
Ellipsis · Coding · Other
Ellipsis is an AI code review tool and cloud coding agent platform that reviews every pull request and runs coding agents from a config file in your repo.
About Ellipsis AI Code Reviews & Bug Fixes
What Is Ellipsis
Ellipsis is a platform for AI software engineers. It started in 2023 as a GitHub app that reviewed code and fixed bugs, and it has grown into managed cloud coding agents that build features, run code review, and look into production issues.
The pitch is simple. You describe an agent in a YAML file inside your repository, and Ellipsis runs it in a cloud sandbox with the permissions and spending limits you set. You can swap the harness between Claude Code, Codex, and Gemini CLI by editing one block of config, so your review and fix workflow is not tied to one model vendor.
The tradeoff is scope. Ellipsis is built for teams that live in GitHub and a few chat tools, it prices on usage instead of a flat seat, and there is no free plan. It also assumes you are fine with config files, since there is no click-and-go review bot for someone who wants zero setup.
Getting Started
- Create an account at ellipsis.dev and install the Ellipsis GitHub app on the repositories you want it to watch.
- Add an agent config in YAML, or pick a template from the gallery, then commit it so a pull request deploys it.
- Set the environment, permissions, secrets, and a budget cap for the agent.
- Open a pull request or mention @ellipsis on an issue, and the agent reviews the push or starts work in a sandbox.
- Follow the session live, send follow-up messages to steer it, then review the diff and the merged changes.
Product Information
A quick look at Ellipsis AI Code Reviews & Bug Fixes's pricing, supported platforms, and performance.
Best for
The users, tasks, and scenarios where this tool fits best.
Users
- Small engineering teams
- Solo developers and indie hackers
- Platform and DevSecOps engineers
- Open source maintainers
Tasks
- Reviewing every pull request automatically with a GitHub code review app that comments on the diff as each push lands.
- Fixing bugs
- Building features from a written task, since the same agents run multi-step work in a sandbox rather than only commenting.
- Investigating production issues after an alert, where the agent digs through logs and code to explain what broke.
Scenarios
- A busy repository where every push needs a review and nobody has time to read the whole diff before it merges.
- A team standardizing on one AI code review workflow across many repositories, with the rules kept in version control.
- Work that should run on a schedule or after a Sentry alert, not only when someone remembers to ask.
Key features
Automated Pull Request Review
Ellipsis reviews each push to a pull request and leaves comments on the lines that matter. You define the reviewers yourself, so the checks follow your standards instead of a generic rule set. This automated code review is the core feature the product launched on.
Agents Defined in YAML
You describe an agent in a config file in your repo: the prompt, environment, permissions, and budget. Because the config lives in version control, an agent changes through a pull request like any other piece of code. That keeps cloud coding agents reviewable and reproducible instead of hidden in a vendor dashboard.
Claude Code, Codex, and Gemini CLI Harnesses
Ellipsis runs sessions on Claude Code, Codex, or Gemini CLI, and you change between them by editing one block of YAML. You keep the same permissions, triggers, and audit trail while you test which model handles your code review and AI bug fixes best. The platform stays the same when the model changes.
Triggers Across GitHub, Slack, Linear, and Sentry
Agents start from a cron schedule, a GitHub event, a Linear issue, a Slack message, or a Sentry alert. You can also mention @ellipsis on GitHub, Slack, or Linear and get an answer in the thread. That turns a review bot into background agents that act without a person in the loop.
API, SDKs, and CLI
Every dashboard action is scriptable through a REST API, with Python and TypeScript SDKs and a CLI. You start a session with one API call and stream its output, which fits teams wiring managed coding agents into their own tooling. Webhooks tell you when a session changes.
Security, Budgets, and Audit Trail
Permissions are set before a session starts, credentials are scoped to the repos you allow, and code lives only in the sandbox before it is deleted. Spending can be capped per session, agent, and account, and every command and file change is logged after the sandbox is gone. Ellipsis states SOC 2 Type I certification, with Type II in progress.
Pros and cons
Pros
- Reviews run on every push, so no commit waits for a human to find the time.
- Agents are defined in YAML inside the repo, which keeps the AI code review workflow version-controlled and easy to reproduce.
- You can move between Claude Code, Codex, and Gemini CLI without rebuilding your setup.
- Usage-based pricing with no per-seat or idle fees suits teams whose review volume moves up and down.
- Spending budgets and a full audit trail give a platform team the controls it needs before rolling agents out.
Cons
- There is no free plan, so a solo developer pays from the first session instead of testing on a small repo for free.
- The workflow assumes comfort with config files and GitHub, since there is no zero-config review bot for teams that want a one-click install.
- Pricing is tokens plus a 10% platform fee, so the monthly cost depends on how much the agents run, which is less predictable than a flat seat.
- Some integrations are still limited: the Sentry trigger is marked as not generally released, so production alert handling may not be ready for every team.
Frequently asked questions
Ellipsis is an AI code review tool and cloud coding agent platform. It reviews each push to a pull request, fixes bugs, builds features, and runs background agents that you define in a YAML config file.
Related content
Explore related tools, skills, and articles for Ellipsis AI Code Reviews & Bug Fixes.
Ellipsis AI Code Reviews & Bug Fixes Alternatives
Base Code
Base44 (Wix) · Coding · ProductivityBase Code is a shared cloud development environment from Base44. Teammates describe changes to an existing GitHub codebase in chat, preview them live and send them back as pull requests.
Google EmbeddingGemma 2
Google DeepMind · Writing · Image · CodingGoogle EmbeddingGemma 2 is a 740M-parameter open model that turns text, code, images, audio, and video frames into vectors in one 768-dimension space, running on phones and laptops.
GitGlow
GitGlow · Coding · ProductivityGitGlow is a local-first desktop Git client that lets independent developers review their own code and their AI agents' changes in one place before shipping.