Name: Applitools
Price: 667 USD

The facts, with their sources

What it does

Writes code: PartlyAutonomous turns a plain-English description into executable test steps, and TestGenAI generates Cypress test code from a recording, but it authors tests rather than application source code.Type what a step should do. An LLM turns your description into a full test step, fixes typos, and breaks complex actions into simpler ones.applitools.com/platform/autonomous
Debugs and fixes: PartlyEyes MCP localises a visual failure to the component file and source line and hands diff coordinates to the coding agent, which then applies the fix; Applitools itself does not land the change.The agent investigates the batch, explains the visual diff in plain English, and applies the code fix straight from your prompt.applitools.com/solutions/agentic-testing
Writes tests: YesAutonomous crawls a site from a URL and builds a starter suite with Visual AI checkpoints already placed, so tests are produced rather than hand-written by the buyer.Enter your homepage URL, or a list of URLs, and Autonomous crawls your sitemap to build a starter test suite with Visual AI checkpoints already in…applitools.com/platform/autonomous

How much it does unattended

Runs autonomously: PartlyTests run unattended on a schedule and in CI and maintenance is grouped automatically, but baseline resolution is deliberately gated behind a human authorisation step.The eyes_resolve_save tool enforces an explicit human-in-the-loop authorization gate, ensuring your team retains complete governance over the visual…applitools.com/solutions/agentic-testing
Agent permissions: YesThe MCP server enforces what a connected agent may settle: review gates on the server side require verified visual evidence before a test or baseline can be resolved.Server-enforced review gates require verified visual evidence before tests or baselines can be resolved.applitools.com/solutions/agentic-testing

Which models it runs on

Claude: YesClaude Code is named as a natively supported agent environment for the Eyes MCP server, set up through standard MCP client configuration.Works natively with leading AI agent environments—including Cursor, Claude Code, GitHub Copilot, and Cline—with zero friction and setup in under two…applitools.com/solutions/agentic-testing
GPT: PartlyGitHub Copilot is listed as a supported agent environment, which implies OpenAI models in the loop, but no page states that GPT models themselves can be selected or configured.Works natively with leading AI agent environments—including Cursor, Claude Code, GitHub Copilot, and Clineapplitools.com/solutions/agentic-testing
Model choice: NoThe checking engine is the vendor's own: a proprietary Deterministic Language Model plus Visual AI trained in-house, so the buyer does not choose the model that does the validation.Applitools has more than a decade of training our algorithms and models for superior accuracy.applitools.com

Where you use it

In your editor: YesVisual AI is delivered inside the editor's agent chat through MCP, with Cursor, Copilot and Cline named, so developers stay in the IDE to triage visual failures.Applitools gives AI coding agents deterministic Visual AI eyes directly inside their chat workflow.applitools.com/solutions/agentic-testing
In your pipeline: YesCI/CD pipeline integration is a Starter-tier line item and the Eyes page names GitHub Actions, GitLab CI and CircleCI for automatic checks on pull requests.Run visual checks on local dev builds and CI/CD pipelines (GitHub Actions, GitLab CI, CircleCI) automatically.applitools.com/platform/eyes
In a browser: YesThere is a hosted SaaS surface with dashboards for results and baselines, offered as the quickest way to start, with private-cloud and on-prem alternatives.Get up and running in minutes using Applitools as a SaaS solution.applitools.com

Whose machine it runs on

Self-hosted: PartlyOn-premises deployment inside your own firewall is offered, but the pricing page lists custom on-premises deployment as an optional Enterprise add-on rather than part of any standard plan.applitools.com/platform-pricing

Building with it

SDK languages: YesThe Eyes SDK picker enumerates bindings across Java, C#, JavaScript/TypeScript, Python and Ruby, bound to specific runners such as Selenium, Playwright, Cypress and WebdriverIO.applitools.com/docs/eyes
Tracing and evals: YesEach run is inspectable: automated maintenance and root-cause analysis ship on Starter, and REST endpoints return batch results and statistics for programmatic reporting.Applitools exposes a set of REST API endpoints that allow you to retrieve data programmatically so that you can access data in the way that best…applitools.com/docs/eyes/reference/serv…

What it costs to run

How it meters: YesBilling is by checkpoint rather than by seat: plans quote component and page checkpoint allowances while users are unlimited on every tier shown.Unlimited users 100,000 component checkpoints or 1,000 page checkpointsapplitools.com/platform-pricing
Cost of one job: YesThe cheapest paid plan is $667 per month billed annually for 100,000 component checkpoints or 1,000 page checkpoints, so one page check costs roughly 67 cents at that allowance.$667 / mo, paid annually Unlimited users 100,000 component checkpoints or 1,000 page checkpointsapplitools.com/platform-pricing
Free tier: PartlyYou can self-register for a free trial without talking to sales, but the only standing plans are paid and no indefinite free tier is offered on the pricing page.You can sign up for a free trial account and explore our documentation and tutorials to get familiar with our platform.applitools.com
API access: YesDocumented REST APIs exist for both Autonomous and Eyes, authenticated with an account API key, covering batch results, statistics, properties and SCM status.Applitools exposes a set of REST API endpoints that allow you to retrieve data programmaticallyapplitools.com/docs/eyes/reference/serv…
MCP server: YesA Visual AI MCP server is included from the Starter tier and is published on the official MCP registry with inspection, resolution and review tool families.Distributed via @applitools/mcp across three distinct families - 10 Inspection tools, 4 Resolution tools, and 3 Review tools on the official Model…applitools.com/solutions/agentic-testing
Bring your own key: PartlyYou can bring your own infrastructure, running Applitools in a dedicated cloud on AWS, Azure or GCP, but the pricing page gates private cloud as a Professional add-on and includes it only on Enterprise.Private cloud & SSOapplitools.com/platform-pricing

Buying it for a team

A company can buy it: YesPlans are sold to organisations, with unlimited users on every tier and named Professional and Enterprise options arranged through sales conversations.Flexible but predictable pricing for teams and projects of all sizesapplitools.com/platform-pricing
Seat model: NoThere is no per-seat charge and no seat floor or ceiling: every published tier states unlimited users, and the metering is on checkpoints instead.applitools.com/platform-pricing
Pooled budget: YesThe checkpoint allowance is an account-level pool drawn on by unlimited users, not a per-person entitlement, so one team can consume another's headroom.applitools.com/platform-pricing
Admin controls: PartlyServer-side gates constrain what a connected agent may resolve and private cloud adds retention and security controls, but no general admin policy console is documented on the pages read.It includes Single Sign-On (SSO/SAML) integration, customizable data retention policies, and enterprise-grade security controls necessary for…applitools.com/platform-pricing
Single sign-on: PartlySSO with SAML is supported, but the pricing page bundles it into a paid Private cloud & SSO add-on on Professional and includes it only on the Enterprise tier.Single sign-on (SSO / SAML) includedapplitools.com/platform-pricing

What happens to your code

Opt out of training: YesThe vendor states flatly that snapshots, component source and DOM data never reach external foundation models or LLM training, with comparisons run in its cloud or a local container.Page snapshots, proprietary component source code, and DOM data are never sent to external foundation models or used for LLM training.applitools.com/solutions/agentic-testing
Data residency: PartlyYou can choose the hosting location by running in a dedicated AWS, Azure or GCP cloud or on your own hardware, but no page names the jurisdictions or regions of the default SaaS.Run Applitools in a dedicated cloud on Amazon, Azure, or GCP to a flexible, secure solution.applitools.com
Getting out: PartlyResults can be pulled out through the REST batch endpoints and private cloud allows custom retention, but nothing read states export of baselines in bulk or whether unused checkpoints carry over.customizable data retention policiesapplitools.com/platform-pricing
Certifications: YesThe vendor claims ISO 27001 and SOC 2 Type II compliance with current penetration testing reports, and publishes GDPR and CCPA pages plus a trust centre.Applitools products are ISO 27001 and SOC 2 Type II compliant and have current pen testing reports.applitools.com

Plans and pricing

Price and terms

Free to start: PartlyPartlySelf-service registration for a free trial account
The terms in full
the vendor does not publish the trial length, its checkpoint allowance or whether a card is required, and no standing $0 plan appears on the pricing page.

Checked 24 October 2026.