Ellipsis vs Greptile: How They Actually Differ

## Ellipsis (/tools/ellipsis)
Tagline: AI agent that reviews every PR — comments on style, bugs, and architecture inside GitHub.
Category: AI Coding Agents
Pricing model: paid
Starts at: Managed SaaS — $0 base + usage (tokens at cost, $0.142 / vCPU-hour, $0.024 / GB-hour, 20% platform fee)
Platforms: web
Public repo: none listed
Docs: none listed
Verified plans (read 2026-08-22): Managed SaaS $0 base + usage (tokens at cost, $0.142 / vCPU-hour, $0.024 / GB-hour, 20% platform fee) | Your AWS Account $0 base + usage + AWS costs (20% platform fee; CPU, memory, and tokens billed by AWS) | Standard support $5,000 / month | Advanced support $7,500 / month | Premier support $15,000 / month
Strengths on file: Functions as a GitHub App that automatically reviews pull requests.; Can generate and commit code fixes directly to a pull request in response to review feedback.; Reviews cover code style, potential bugs, and architectural concerns.; Can produce pull request descriptions generated from code diffs.
Limits on file: Supports only GitHub, with no integration for GitLab, Bitbucket, or other Git platforms.; Free tier imposes usage limits consistent with its freemium pricing.
Ellipsis is an AI code-review agent that reviews every pull request inside GitHub. It runs as a GitHub App that reviews pull requests automatically.
Its reviews cover code style, potential bugs, and architectural concerns, and it can go further than commenting: it can generate and commit code fixes directly to a pull request in response to feedback, and produce pull request descriptions from code diffs.
It fits teams that work in GitHub and want automated PR review and fixes. Caveats: it supports only GitHub, with no integration for GitLab, Bitbucket, or other Git platforms; there is no IDE extension or CLI, so interaction is limited to the GitHub web interface; and the free tier imposes usage limits.
Among code-review tools, it is a GitHub-native agent that both comments and commits fixes rather than a standalone editor or CLI reviewer.

## Greptile (/tools/greptile)
Tagline: AI-powered code review that comments on PRs by understanding the entire codebase, not just the diff.
Category: AI Coding Agents
Pricing model: paid
Starts at: Pro — $30 /seat/month
Platforms: web, cli
Public repo: none listed
Docs: none listed
Verified plans (read 2026-08-21): Starter Free | Pro $30 /seat/month | Enterprise Custom pricing
Strengths on file: Analyzes the entire repository context rather than only the changed lines in a pull request.; Posts review comments directly on pull requests in the connected Git host.; Integrates with GitHub (and GitLab) repositories to ingest and review code.; Uses large language models to reason about cross-file relationships and project conventions during review.
Limits on file: No free tier is offered; access requires a paid plan.; Code must be hosted on a supported Git provider, creating dependency on that integration.
Greptile is an AI code reviewer that comments on pull requests by reading the whole codebase, not just the diff.
Its edge is whole-repo context: it analyzes the entire repository rather than only the changed lines, reasoning about cross-file relationships and project conventions, and posts its comments straight onto PRs in your Git host (GitHub and GitLab).
It's for teams that want reviews aware of the broader codebase, not line-by-line diff checks. Caveats: there's no free tier — it's paid only — review runs only through the web platform with no IDE plugin or CLI, and your code has to live on a supported Git provider.
It's a PR-review tool like CodeRabbit, distinct from the in-editor assistants.