Holmes vs Greptile
They do different jobs, and they work well together. Greptile reads your whole codebase to review each change and can be self-hosted. Its TREX beta also runs pull request branches in a sandbox. Holmes tests the running product on pull requests, deploys and schedules, across web, mobile and APIs, and reports in Slack.
Greptile reviews code against a graph of your codebase. Holmes opens your app in a real browser and works through each journey from a plain-language goal, so it finds what only breaks when someone uses the product.
Greptile's reviews and TREX runs happen on pull requests. Holmes also tests on every deploy, on a schedule or by webhook, so it checks the product your customers use, not only the branch.
Holmes tests native Android and iOS apps on emulators and simulators and checks every request behind a journey in context. Greptile's pages don't describe mobile or API testing.
Holmes sends findings to the pull request, the ticket, Slack, Microsoft Teams and email, with a recording, screenshots and logs. Every plan has unlimited users, so product and support see them too.
Based on each product's public pages on 17 November 2026. A red cross means Greptile's pages don't describe it, not that it can't be done.
| Capability | Holmes | Greptile |
|---|---|---|
| What it checks | ||
| Main job | Yes: Tests web apps, mobile apps and APIs from plain-language goals | Yes: Reviews every pull request with a graph of the whole codebase[1][2] |
| Code review | Not covered: Not covered. Holmes reads the code change to decide what to test | Yes: Comments with suggested fixes, learned from your team's past reviews[1][2] |
| Uses the app like a person would | Yes: Plain-language goals, run in a real browser | Partly: TREX, in beta: starts the pull request branch in a sandbox and clicks through the flows it affects[1][4] |
| Finds bugs nobody scripted | Yes: Agents explore and test paths nobody wrote down | Partly: TREX works out which user flows a change affects; custom flows are written step by step[4][7] |
| Native Android apps | Yes: Native Android apps on an Android emulator | Not described: Not on its pages. Native mobile app testing |
| APIs | Yes: Every request behind a journey, checked in context | Not described: Not on its pages. API testing |
| Accessibility | Yes: WCAG audits in a real browser, with keyboard navigation | Not described: Not on its pages. WCAG audits |
| Workflow | ||
| Pull requests | Yes: Check and comment on GitHub and Bitbucket pull requests. GitLab supported | Yes: Reviews on GitHub, GitLab, Bitbucket and Cursor Origin; Enterprise adds self-managed GitHub, GitLab and Bitbucket, Gitea and Perforce. TREX runs on GitHub and GitLab[1][5] |
| Deploys and schedules | Yes: On every deploy, on a schedule, or by webhook | Partly: Runs on pull requests. Testing after a deploy or on a schedule isn't on its pages[4][7] |
| Context it reads | Yes: Reads the code change, ticket, spec (Notion, Confluence), design (Figma) and analytics (PostHog) | Partly: The whole codebase, past review comments, and Jira, Linear and Notion on Pro. Figma or analytics aren't on its pages[1][3][8] |
| Coding agents (MCP) | Yes: Start runs and read findings from Claude and Cursor via MCP | Yes: MCP server, and fixes sent to Claude Code, Codex, Cursor or Devin[1][2] |
| Evidence with each bug | Yes: Recording, screenshots, console and network logs, steps to reproduce | Partly: Code comments with suggested fixes; TREX adds a captioned video of each flow. Console and network logs aren't described[2][6] |
| Where findings go | Yes: Findings on the pull request, the ticket, Slack, Microsoft Teams and email | Partly: On the pull request, inline on the diff. Slack or Microsoft Teams alerts aren't on its pages[2][6] |
| Pricing and security | ||
| What you pay for | Yes: Priced on how much Holmes tests, never on seats | Note: Credits: 1 to 10 per review, 3 per TREX run[3][4] |
| Users | Yes: Unlimited users on every plan | Note: One developer on Starter; unlimited developers on Pro, no seat fees[3] |
| Entry price | Yes: Free plan. Paid plans from €199 ($219) a month | Note: Starter free with 50 credits a month; Pro $1 a credit; Enterprise on request. Free for qualifying open-source projects[3] |
| Getting started | Yes: €250 in free credits, no credit card | Yes: 14-day free trial of Pro[3] |
| Self-hosting | Not covered: Not covered. Cloud only | Yes: On Enterprise, on-premises or in your VPC[3][9] |
| AI and your data | Yes: Zero-data-retention AI processing. Source code is never stored | Note: Zero-data-retention agreements with its LLM providers; code stored encrypted on AWS. De-identified data may train models unless you opt out. SOC 2 Type II[9] |
| Company base | Yes: EU company (Ghent, Belgium), GDPR compliant | Note: San Francisco, United States[10] |
Holmes sells QA software. Greptile sells AI code review, so we overlap mainly where its TREX beta tests pull requests. We've tried to be fair: every statement about Greptile comes from its own public pages, linked below with the date we checked them. Products change. If something here is out of date, email support@useholmes.com and we'll fix it.
AI agents that review and test pull requests with full context of the codebase.Checked 17 November 2026.
Greptile is an AI code review agent that automatically reviews every pull request.Checked 17 November 2026.
Unlimited developers and agents, no seat fees.Checked 17 November 2026.
Uses the app like a user would: clicking, typing, and observing what's on the screen.Checked 17 November 2026.
Simulation supports GitHub and GitLab repositories.Checked 17 November 2026.
Recordings and other evidence are kept for 30 days.Checked 17 November 2026.
TREX runs it as written whenever a pull request touches it.Checked 17 November 2026.
Greptile finds Jira tickets related to your pull requests and reviews the code against the requirements.Checked 17 November 2026.
Self-hosted Greptile runs in your own infrastructure, on-premises or in your virtual private cloud (VPC), so your code never leaves your environment. Greptile is SOC 2 Type II compliant.Checked 17 November 2026.
We are a small group of researchers, engineers, and designers in San Francisco.Checked 17 November 2026.
They do different jobs. Greptile is an AI code reviewer: it reads the whole codebase and comments on each pull request. Holmes is a QA platform: it tests the running product the way customers use it, on web, mobile and APIs. Greptile reviews the code; Holmes checks that the product still works.[1][2]
Many teams will. Code review catches problems in the code before it merges. Holmes catches what only shows up when someone uses the product, on pull requests and again after every deploy or on a schedule. Holmes does not review code, so it does not replace Greptile.
TREX is Greptile's end-to-end testing agent, labelled beta on its homepage. It starts the pull request branch in a sandbox, uses the app like a user would, and posts what it tested on the pull request with a video of each flow. It runs on GitHub and GitLab repositories. Holmes also tests after deploys and on schedules, covers native mobile apps and APIs, reads specs, designs and analytics, and sends findings to Slack, Microsoft Teams and the ticket.[1][4][5][6]
Greptile charges credits. A review costs 1 to 10 credits and a TREX run 3. Starter is free for one developer with 50 credits a month, Pro is $1 a credit with no seat fees, and Enterprise is priced on request. Holmes has a Free plan, Starter at €199 a month for up to 3 products, and Team from €499 a month, all with unlimited users. Holmes is priced on how much it tests, never on seats.[3][4]
If what you need is AI review of every pull request, Greptile does that job and Holmes does not. It also offers self-hosting. If you need to know that sign-up, checkout and your other journeys still work for customers, after every deploy as well as on pull requests, try Holmes alongside it.[1][3][9]
Evaluating Greptile? See what Holmes finds first.
Free, no scripts, no SDK, and no test account needed to start.