Ellipsis AI

Ellipsis is an AI-powered developer platform designed to automate software engineering work, with a strong focus on code review, bug detection, code generation, and cloud-based coding agents. The product originally gained attention as an AI code-review tool that could automatically analyze GitHub pull requests, identify potential bugs, suggest improvements, and answer engineering questions. It has since expanded into a broader platform for running and governing AI coding agents in cloud environments.

Ellipsis was started in 2023 and went through Y Combinator’s Winter 2024 batch. The company says it launched with automated code review before expanding into a platform of AI agents that can work across development workflows. Its current platform is designed to let engineering teams deploy agents that operate from GitHub, Slack, Linear, schedules, and other triggers.

AI code review

One of Ellipsis’s core capabilities is automated pull-request review. Ellipsis can review code as commits are pushed to a pull request and provide inline findings on potential defects. Rather than repeatedly analyzing an entire pull request, its current review system focuses on newly added commits after the previous successful review.

The built-in reviewer is designed to look for issues such as logic errors, edge cases, error handling problems, concurrency issues, security concerns, regressions, and data or deployment hazards. Ellipsis emphasizes confirmed defects rather than generating large numbers of speculative style comments.

Teams can customize the review process using a code_review.yaml configuration file. This allows organizations to define reviewers, filters, budgets, permissions, pull-request scope, and other review behavior. Teams can also create gatekeeper agents to filter findings before they are posted.

AI code generation and bug fixing

Ellipsis can work as an AI engineering teammate rather than simply a passive reviewer. Developers can assign work through GitHub comments and allow the system to generate working, tested code.

The supplied Ellipsis page highlights async code generation, where developers can assign tasks through GitHub comments and receive completed code later. It also describes automated bug fixing, release-note generation, codebase questions, and build-issue assistance.

The current platform expands these capabilities through cloud coding agents. Agents can operate inside isolated sandboxes with repositories, tools, credentials, budgets, and environment settings defined for each session. This allows development work to happen without requiring every agent to run directly on a developer’s computer.

AI agents

Ellipsis now positions itself as infrastructure for software-development agents. Agents can be defined as code using YAML and deployed through Git-based workflows.

Teams can create agents for activities such as code review, dependency updates, standup summaries, migration checks, release-note generation, and other repetitive engineering tasks. Agents can be triggered by GitHub events, Slack mentions, Linear issues, schedules, and other supported events.

The platform also supports multiple coding agents and models, including Claude, Codex, GitHub Copilot, Gemini, Cursor, Grok Code, OpenCode, and others, although some integrations are marked as coming soon on the current website.

Cloud development environments

Each Ellipsis agent session can run in its own isolated cloud sandbox with the relevant repository and development toolchain available. This allows agents to inspect code, install dependencies, execute development tasks, run tests, and work with project-specific environments.

The platform provides controls for credentials, permissions, budgets, and session visibility. Organizations can also choose between Ellipsis’s managed cloud environment and deployments within their own AWS infrastructure.

Automation

Automation is a major part of Ellipsis’s current platform. Engineering teams can define agents that start automatically when specific events occur.

For example, an organization can configure an agent to review pull requests, summarize merged work on a schedule, investigate development issues, or perform recurring maintenance tasks. This moves AI assistance from a manually invoked chatbot toward event-driven engineering automation.

Style guides and team feedback

Ellipsis supports natural-language style guides that allow teams to describe their coding standards and have the AI identify violations during reviews. The original product also emphasized learning from team feedback so that reviews could become more aligned with the types of comments developers consider useful.

This can help organizations maintain consistent engineering practices across repositories without requiring every developer to manually enforce the same standards.

Questions and developer assistance

Developers can interact with Ellipsis from supported collaboration tools and ask questions about their code. The platform can use repository context when responding, making it more useful for engineering questions than a general-purpose AI assistant without access to the project’s source code.

The current platform supports mentions through GitHub, Slack, and Linear, allowing agents to respond where engineering work is already happening.

Release notes and engineering reporting

Ellipsis can also automate communication around software development. Its original product included automatic change-log generation and weekly summaries explaining what changed in a codebase and why the changes matter.

The current platform extends this concept through scheduled agents and examples such as standup automation, where an agent can summarize engineering activity and post the result to Slack.

Security and governance

Security is an important part of Ellipsis’s positioning. The supplied site states that Ellipsis is SOC 2 Type I certified and that source code is not permanently retained. The site explains that source code exists on its servers only while a job is being processed in a private AWS VPC.

The current platform adds scoped credentials, budget limits, auditability, isolated sessions, and deployment options. Enterprise customers can use their own AWS environment, bring their own encryption keys, and apply additional governance controls.

Ellipsis also states that its automated review system does not approve pull requests, request changes, push commits, or merge code. This keeps the review function separate from final human approval.

Pricing

The supplied Ellipsis Revamp page lists a Pay Per Seat model with unlimited usage at $20 per developer per month and a 7-day free trial with no credit card required.

However, Ellipsis’s current official pricing page has since changed to a usage-based model for its broader cloud coding-agent platform. The current managed SaaS pricing is based on LLM token costs plus a 10% platform fee, with no per-seat fees. CPU and memory usage are also billed according to consumption. The same platform can alternatively be deployed into a customer’s AWS account, where AWS costs apply.

The current platform includes a $100 starting credit and does not use traditional feature-gated tiers. Enterprise support packages are available separately, starting at $5,000 per month for Standard support, with Advanced and Premier packages priced at $7,500 and $15,000 per month respectively.

Because the supplied Revamp page and current official pricing page describe different pricing models, the current Ellipsis.dev pricing should be treated as the source of truth for present-day pricing.

Review

Ellipsis is a strong AI development platform for teams that want to automate repetitive software-engineering tasks rather than use AI solely for autocomplete or conversational coding.

Its strongest capability is the combination of code-aware AI review and autonomous development workflows. The code-review system analyzes changes in repository context and focuses on confirmed defects, while the broader agent platform allows teams to run AI coding agents in isolated environments with controlled permissions and budgets.

The GitHub integration is particularly useful because code review and development work can happen directly within existing pull-request workflows. Slack and Linear integrations further extend the agents into the tools engineering teams already use.

Another advantage is customization. Teams can define agents and review pipelines using configuration files, control their permissions, set spending limits, and create specialized workflows for different engineering tasks.

The main limitation is that Ellipsis is specialized for software development. It is not a general-purpose AI assistant, and its value is highest for organizations with active software repositories and established engineering workflows. The current usage-based pricing can also make costs less predictable than the older $20-per-developer model because spending depends on model usage and compute consumption.

Ellipsis is best suited to software engineering teams, DevOps organizations, platform teams, startups, and enterprises that want AI agents to handle code review, development automation, debugging, maintenance, and engineering reporting while keeping human developers in control of important decisions.