> ## Content Index > Fetch the complete content index at: https://www.theaidirectory.com.au/llms.txt > Use this file to discover other available public pages before exploring further. # Ellipsis Review: AI Code Review That Fixes Its Own Bugs (Australia 2026) - URL: https://www.theaidirectory.com.au/review-ellipsis-australia-2026/ - Published: 2026-10-08T23:00:00.000Z - Updated: 2026-10-09T06:29:47.000Z - Description: Ellipsis doesn't just flag bugs — it generates working code fixes as pull request comments, so developers can apply or discard them in one click. The flat US$20 per developer per month plan we scored in June is gone: since 2026 Ellipsis bills usage (tokens at cost plus a 10% platform fee, no seat fe - Author: Adam - Tags: Coding & Development, AI Reviews, #review **13**/25 ★★★☆☆ Good Ellipsis (New York City, United States — seed stage) Verified 9 Oct 2026 Value for Money**3**/5 Ease of Use**4**/5 Feature Depth**3**/5 Australian Fit**2**/5 Reliability & Support**1**/5 Pricing Pricing on application — see the pricing table History Edition 2\. Was [14/25, Jun 2026](https://www.theaidirectory.com.au/review-ellipsis-australia-2026-06/) **The gate:** Five stars need 21+ and Australian Fit of at least 4/5. [Skip to the verdict ](#verdict) *Pricing figures updated 6 September 2026: USD prices are converted at the current rate (1 USD ≈ 1.40 AUD, verified 5 September 2026) and shown ex-GST, replacing the retired 1.55 + GST basis. A GST-registered Australian business that supplies its ABN is not charged GST on an imported digital service; other buyers should allow \~10% on top (*[*Methodology*](https://www.theaidirectory.com.au/methodology/)*).* > **Verdict (TL;DR):** Ellipsis is the most action-oriented AI code review tool in this category. When it finds a bug or style violation, it doesn't just describe the problem — it generates a working code fix as a PR comment, ready to apply. The flat US$20 per developer per month pricing we scored in June no longer exists: as at October 2026 Ellipsis charges usage — tokens at cost plus a 10% platform fee, with no per-seat fee — and positions itself as a cloud coding-agent platform rather than a review tool, so this edition updates the price and flags a full re-review. The significant limitations are that it is GitHub-only, has no self-hosted deployment, is a small seed-stage company, and only began GitLab support in beta as of 2026\. For Australian dev teams comfortable with GitHub and US cloud infrastructure, it is worth serious consideration at its price point. > > **Score: 13/25 — ⭐⭐⭐ Good** > > *Last reviewed: 26 June 2026 by Adam.* --- > **Earlier editions of this review:** [26 Jun 2026 — 14/25](https://www.theaidirectory.com.au/review-ellipsis-australia-2026-06/). This is edition 2, dated 9 Oct 2026\. We keep every edition so you can see how the tool and our verdict have moved — [how and why](https://www.theaidirectory.com.au/review-timeline/). ## What Ellipsis is Ellipsis is an AI code review tool with a defining feature: when it identifies a bug, antipattern, or style violation, it generates a working code fix — not just a description of the problem. That fix appears as a PR comment with a code block the developer can review, discuss, and apply with a single click. The company was founded in 2023 by Hunter Brooks and Nick Bradford, based in New York City. It is seed stage, having raised US$2 million in a funding round in April 2024\. That small funding base is relevant context for Australian businesses evaluating long-term vendor risk. Beyond per-commit code review and auto-fixing, Ellipsis operates as what it calls an AI teammate. Developers can assign tasks to Ellipsis directly via GitHub comments — "@ellipsis, fix this" — and Ellipsis will generate and open a pull request with the working implementation. It also creates release notes, weekly codebase summaries, and Slack reports on repository activity. The tool integrates with GitHub (fully supported) and GitLab (beta as of 2026). Bitbucket and Azure DevOps are not supported. --- ## What it costs in Australia Ellipsis has changed its pricing model entirely. Read in a browser on an Australian connection on 9 October 2026, `ellipsis.dev/pricing` is headed "Tokens + 10%": there is no per-seat fee, you pay the model tokens your agents consume at cost plus a 10% platform fee, and on the managed service you also pay for compute. The June edition of this review scored a flat US$20 per developer per month Business plan with a free tier for public repositories; neither appears on the page any more. | Component (as at 9 Oct 2026) | Managed SaaS (USD) | Your own AWS account | | --------------------------------- | ------------------------------------------------------------------------------------------- | -------------------- | | Per-seat fee | $0 | $0 | | Model tokens | At cost, no markup | Billed to you by AWS | | Platform fee | 10% of token cost | 10% of token cost | | Compute | $0.142 per vCPU-hour; $0.024 per GB-hour of memory (\~A$0.20 and \~A$0.035 at 1.44, ex-GST) | Billed to you by AWS | | Support packages (optional) | Standard $5,000, Advanced $10,000, Premier $15,000 a month | Same | | June 2026 edition, for the record | Free (public repos) / Business US$20 per developer per month (\~A$28) / Enterprise custom | — | Both offerings carry the full feature set — "no feature gates, no tiers" — and interactive sessions such as @ellipsis in Slack can run on your own Claude Code or Codex subscription instead. The support packages are optional: the platform is self-serve. What you pay now depends on how many tokens your agents burn, so a team cannot read a monthly figure off the page the way it could in June. Enterprise adds SAML SSO, priority support, and SLA guarantees — pricing on request. The June edition recorded public repositories as free with no account; the current page's FAQ asks "Is it free for open source?" but the answer did not render to our read, so treat that as unconfirmed. The comparison that made Ellipsis the cheapest paid option in this category (US$20 a month against CodeRabbit Essentials at US$24, Greptile Pro at US$30 and Graphite Team at US$40) no longer holds in any simple form: a light user may pay far less than US$20 under usage billing, a heavy user far more. For small Australian dev teams that valued predictable billing, that predictability is the thing that has gone. **GST note:** Ellipsis is a US seed-stage company with no Australian GST registration. AUD figures use approximately AU$1 = US$0.71 (1.40 exchange rate) ex-GST (verify at publication). --- ## How it works in practice Install Ellipsis as a GitHub App on your organisation. Every subsequent commit on every pull request triggers a review — Ellipsis reviews per-commit, not just when the PR is opened, which means fast-moving PRs get continuous feedback rather than a single pass at creation. Each review appears as a PR comment from the Ellipsis bot. It flags: - **Logical bugs** — incorrect logic, missing edge cases, potential crashes - **Style guide violations** — based on your configured style guide, which you define in natural language or by providing examples - **Antipatterns** — code that will work but will cause problems at scale or with future maintainers For each finding, Ellipsis generates a working fix in the same comment. You can apply the fix, ask Ellipsis to revise it ("Make this approach more defensive"), or dismiss it. The auto-fix generation is the standout feature — it turns code review comments from tasks on a to-do list into one-click decisions. The task assignment feature works by commenting on a PR or issue: "@ellipsis, implement the retry logic for this API call." Ellipsis processes the request, writes the code, and opens a pull request. Teams report this is best suited to well-scoped, lower-risk tasks rather than significant architectural work. The reporting feature sends daily, weekly, or monthly summaries of codebase activity to email or Slack — useful for engineering managers who want visibility without reading every PR. --- ## What it does well **Auto-generated fixes, not just comments.** This is the genuinely differentiated feature. When Ellipsis finds a problem, it shows you the fix. The developer decides in one click — apply, modify, or dismiss. This reduces the back-and-forth between reviewer and developer and is why Ellipsis-using teams reportedly merge PRs 13% faster on average. **Per-commit reviews.** Reviewing every commit on every PR, rather than just the initial PR opening, means issues raised during active development are caught sooner. On a PR where a developer makes five commits while implementing feedback, Ellipsis reviews all five. **No per-seat fee, tokens at cost.** The October 2026 model removes seat fees altogether and passes model tokens through at cost plus 10%, which suits teams that want to pay only for what their agents do. The flat US$20 plan that made billing predictable in June is gone, so this strength has changed shape rather than disappeared. **Codebase reports.** The weekly repository summaries and activity reports are a lightweight product analytics feature that larger tools charge extra for. For engineering managers, getting a plain-English summary of what changed and why — delivered to Slack on Monday morning — has genuine value. **Task assignment via comments.** For well-scoped, lower-risk implementation tasks, being able to direct Ellipsis to write code via a GitHub comment is a productivity tool. It works best for self-contained tasks with clear requirements. --- ## What it doesn't do well **GitHub-only (GitLab still in beta).** Ellipsis's GitLab integration was in beta as of June 2026\. Bitbucket and Azure DevOps are not supported. For Australian enterprises on Azure DevOps, or any team not on GitHub, Ellipsis is not an option. **No self-hosted deployment.** Ellipsis is cloud-only. Source code is sent to Ellipsis's AWS infrastructure in the US for processing, then deleted. LLM requests are routed through OpenAI or Azure OpenAI APIs (also US). Ellipsis explicitly states it has no self-hosting option — "unsuitable for highly isolated environments" per its own documentation. For Australian regulated industries, this is a disqualifier. **Seed-stage company risk.** Ellipsis has raised US$2 million. In a category where CodeRabbit is well-funded and Greptile has raised US$29 million, Ellipsis's funding is thin. This is not a commentary on the product quality — which is genuine — but on longevity risk for Australian businesses evaluating a multi-year tool commitment. Factor this into your decision. **LLM request logging.** Ellipsis uses PromptLayer to log LLM requests for production support purposes. This means your code, as part of LLM prompts, is logged by a third-party service (PromptLayer) as well as Ellipsis itself and the underlying LLM provider (OpenAI/Azure OpenAI). You can opt out via the Settings page, but it is opt-out rather than opt-in. **Narrower review scope than full-codebase tools.** Ellipsis reviews the diff and PR context. It does not index the full codebase the way Greptile does, which limits its ability to catch cross-file issues that require understanding the full architecture. **GitHub Copilot overlap.** If your team already pays for GitHub Copilot, some of Ellipsis's functionality — auto-complete, basic code suggestions — overlaps. The auto-fixing feature is still differentiated, but the full value proposition narrows if Copilot is already in the stack. --- ## Australian fit — code and IP data residency **What happens to your code:** When Ellipsis reviews a pull request, it temporarily checks out the code in a private AWS VPC (US-based). Code is automatically deleted when the review workflow completes and is never persisted between sessions. Ellipsis uses OpenAI or Azure OpenAI APIs to generate reviews and fixes — these requests also route through US infrastructure. **Third-party logging:** LLM requests are logged via PromptLayer by default. This means code snippets included in LLM prompts may be retained by PromptLayer (a US company) for support purposes. Australian businesses should opt out of this logging (available in Settings) and verify the data handling implications independently. **Privacy Act and APP 8:** The same overseas disclosure analysis applies as for other US-hosted code review tools. AWS data processing, OpenAI API calls, and PromptLayer logging all represent overseas disclosures under Australian privacy law. For commercial teams in non-regulated industries, these are manageable with appropriate contractual controls and opt-outs. For regulated industries, the multi-party data flow is harder to control. **No self-hosted option:** Ellipsis's own documentation is explicit — it has no self-hosting option and is not suitable for isolated environments. Any Australian business with data residency requirements cannot deploy Ellipsis in a compliant configuration. **SOC 2 Type I:** Ellipsis holds SOC 2 Type I certification (note: Type I, not the more rigorous Type II). This is a positive but is a lower bar than the SOC 2 Type II certifications held by CodeRabbit and Greptile. Australian Fit score: 2/5\. Cloud-only US infrastructure, GitHub-only (GitLab in beta), no self-hosted option, and multiple third-party data recipients (OpenAI/Azure OpenAI, PromptLayer) create a complex data flow. The logging opt-out, zero code retention policy, and SOC 2 Type I certification provide some assurance. Not suitable for regulated industries. --- ## Alternatives - **CodeRabbit** ([review](https://www.theaidirectory.com.au/review-coderabbit-australia-2026/)) — supports GitHub, GitLab, Bitbucket, and Azure DevOps; 40+ linters; no auto-fix generation; US$24/month annual. Better platform coverage and more mature company; less differentiated on fixing. - **Greptile** ([review](https://www.theaidirectory.com.au/review-greptile-australia-2026/)) — full codebase indexing; self-hosted enterprise option available; US$30/month base with 50-review cap. Higher accuracy and data residency option, but no auto-fixing and variable billing. - **GitHub Copilot** ([review](https://www.theaidirectory.com.au/github-copilot-review-australia-2026/)) — integrated into GitHub natively; inline suggestions and basic code review; stronger enterprise data residency options through Microsoft. If you are already paying for Copilot, compare its code review feature directly to Ellipsis's value proposition before subscribing to both. --- ## Bottom line Ellipsis makes one genuinely differentiated bet: that the most useful thing an AI code reviewer can do is generate the fix, not just describe the problem. That bet is well-executed. The per-commit review cadence and the auto-fix generation are solid product decisions; the flat US$20 pricing that was the third leg has been replaced by usage billing. The concerns are real. It is GitHub-only (with GitLab in beta). There is no self-hosted option. The funding is thin for a category that is consolidating around better-resourced competitors. And the default LLM request logging via PromptLayer adds a data flow complexity that Australian businesses should understand and opt out of. If you are a small-to-medium Australian dev team on GitHub, not in a regulated industry, and willing to back an early-stage product, Ellipsis — now priced on token usage plus 10%, with no seat fee — is worth a trial. The 7-day free trial requires no credit card. Use it on a real codebase before committing. For teams that need platform flexibility, better data residency, or a more established vendor, CodeRabbit or Greptile are the stronger choices. --- ## Score | Criterion | Score | Notes | | --------------------- | --------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | Value for Money | 3/5 | Down from 4 on 9 Oct 2026: the flat US$20 seat that made it the cheapest paid option is gone; tokens at cost plus 10% with no seat fee is fair, but a team can no longer budget from the page, and support packages start at US$5,000 a month | | Ease of Use | 4/5 | GitHub App install is simple; per-commit reviews require no configuration; auto-fix and task assignment are well-designed interactions | | Feature Depth | 3/5 | Auto-fix generation (unique); per-commit review cadence; codebase reports; task assignment; narrower review scope than full-codebase tools; GitLab still in beta | | Australian Fit | 2/5 | GitHub-only (GitLab beta); no self-hosted option; US infrastructure; multiple third-party data recipients; default LLM logging via PromptLayer (opt-out available); SOC 2 Type I only | | Reliability & Support | 1/5 | Seed-stage company with US$2M raised; limited support resources; SOC 2 Type I (not Type II); longevity risk is a genuine concern for multi-year commitments | | **Total** | **13/25** | **⭐⭐⭐ Good** | *Australian Fit gate: Five stars require AU Fit ≥ 4/5\. Ellipsis scores 2/5 on AU Fit — US-only infrastructure, no self-hosted option, GitHub-only platform, and multiple third-party data recipients. Rating sits in the Good (⭐⭐⭐) band on raw score regardless. See our* [*Methodology*](https://www.theaidirectory.com.au/methodology/)*.* --- Update history - **9 Oct 2026** — **New edition: Ellipsis dropped its flat seat price and repositioned as a cloud coding-agent platform. Value for Money 4 → 3, total 14 → 13/25.** Was: Free for public repos; Business US$20 per developer per month, no caps; Enterprise custom. Now, from `ellipsis.dev/pricing` read in a browser on an Australian connection: **per-seat fee US$0; model tokens at cost plus a 10% platform fee; on the managed service compute at US$0.142 per vCPU-hour and US$0.024 per GB-hour; or run it in your own AWS account; optional support packages US$5,000 / 10,000 / 15,000 a month.** The page now describes agents, sessions, environments and budgets for background coding agents rather than the PR-review tool this review scored, so **a full re-review is owed** (logged in Review-Queue). Bare `$`, no AUD; `aud_native: false` stands. Source: AUD queue #480; capture `Dossiers/captures/ellipsis-pricing-2026-10-09.txt`. The June edition is kept unchanged and linked above. ## Frequently asked questions ### What makes Ellipsis different from other AI code review tools? Ellipsis generates working code fixes alongside its review comments, not just descriptions of the problem. When it finds a bug or style violation, it posts a corrected code block in the PR comment that the developer can apply, modify, or dismiss in one click. It also reviews every commit on every PR (not just when the PR is opened), accepts task assignments via GitHub comments, and generates codebase activity reports for Slack or email. The flat US$20/developer/month pricing with no usage caps is also a differentiator in a category where per-review overages are common. ### Is Ellipsis safe for Australian businesses from a data privacy perspective? Ellipsis processes code on AWS infrastructure in the US and routes LLM requests through OpenAI or Azure OpenAI (also US). Code is deleted after each review workflow. By default, LLM requests are logged via a third-party service called PromptLayer — Australian businesses should opt out of this in Settings. Ellipsis holds SOC 2 Type I certification. There is no self-hosted option and no Australian data residency. For commercial teams in non-regulated industries, this is manageable with appropriate controls. For government, defence, health, or finance organisations with strict data residency requirements, Ellipsis is not suitable. ### Is Ellipsis a safe long-term vendor choice? Ellipsis is a seed-stage company that has raised US$2 million as of mid-2026 — significantly less than competitors CodeRabbit or Greptile. The product quality is genuine but the company's longevity risk is higher than established vendors. For month-to-month subscriptions or short-term projects, this risk is manageable. For teams making a multi-year infrastructure decision, consider the vendor stability question carefully alongside the product fit. The 7-day free trial (no credit card) is a low-risk way to evaluate before committing. --- ## Sources - Ellipsis homepage: [https://www.ellipsis.dev/](https://www.ellipsis.dev/?ref=theaidirectory.com.au) \[PRIMARY\] — [Wayback 10 June 2026](https://web.archive.org/web/20260610033509/https://www.ellipsis.dev/) - Ellipsis Code Review documentation: [https://docs.ellipsis.dev/features/code-review](https://docs.ellipsis.dev/features/code-review?ref=theaidirectory.com.au) \[PRIMARY\] — [Wayback 18 February 2026](https://web.archive.org/web/20260218013759/https://docs.ellipsis.dev/features/code-review) - Ellipsis Security documentation: [https://docs.ellipsis.dev/security](https://docs.ellipsis.dev/security?ref=theaidirectory.com.au) \[PRIMARY\] - Ellipsis FAQ: [https://docs.ellipsis.dev/faq](https://docs.ellipsis.dev/faq?ref=theaidirectory.com.au) \[PRIMARY\] — [Wayback 12 March 2026](https://web.archive.org/web/20260312142239/https://docs.ellipsis.dev/faq) - Ellipsis Seed Round announcement: [https://www.ellipsis.dev/blog/ellipsis-raises-a-2m-seed-round](https://www.ellipsis.dev/blog/ellipsis-raises-a-2m-seed-round?ref=theaidirectory.com.au) \[PRIMARY\] — [Wayback 14 September 2025](https://web.archive.org/web/20250914045909/https://www.ellipsis.dev/blog/ellipsis-raises-a-2m-seed-round) - Ellipsis GitLab beta: [https://app.ellipsis.dev/gitlab](https://app.ellipsis.dev/gitlab?ref=theaidirectory.com.au) \[PRIMARY\] — [Wayback 3 September 2026](https://web.archive.org/web/20260903154136/https://app.ellipsis.dev/gitlab) - Ellipsis Codebase Reports: [https://docs.ellipsis.dev/features/reports](https://docs.ellipsis.dev/features/reports?ref=theaidirectory.com.au) \[PRIMARY\] — [Wayback 15 June 2026](https://web.archive.org/web/20260615051029/https://docs.ellipsis.dev/features/reports) - AUD/USD exchange rate: approximately 1.40 as of September 2026 (verify at publication) --- --- *The AI Directory Australia is an independent editorial publication. We don't accept payment for positive reviews. Read our* [*Methodology*](https://www.theaidirectory.com.au/methodology/) *and* [*Affiliate Disclosure*](https://www.theaidirectory.com.au/affiliate-disclosure/)*.*