Docket

What do they actually do

Docket lets teams create and run end‑to‑end web tests in a real browser. You can describe a flow in plain English or record it by clicking through your app, and optionally add “agent rules” to guide edge‑case behavior Quickstart, Test Suites, Welcome.

An AI agent executes the steps by visually locating and interacting with elements (vision‑first) rather than relying only on brittle DOM selectors, and returns structured failure reports with screenshots, console/network traces, and clear reproduction steps Welcome, product blog.

Tests can be run manually, on a schedule (including cron), or from CI/CD. Docket integrates with GitHub to report pass/fail status, can block deploys when checks fail, and can point tests at preview environments (e.g., Vercel previews) via URL overrides Running a Test, CI Integration.

Who are their target customer(s)

• Frontend engineers / QA engineers: They spend time chasing flaky tests and reproducing bugs when UI changes break selectors; they need reliable runs and clear failure artifacts to speed triage and fixes Welcome, product blog.
• Product managers and non‑engineers who need to verify features: They can’t maintain code‑based tests and rely on engineers for validation; they need to author or record checks without writing code Quickstart, Welcome, YC.
• DevOps / release engineers: They need automated release gates but setting up stable, CI‑integrated end‑to‑end checks is time‑consuming; they need simple CI hooks, status reporting, and deployment blocking CI Integration.
• QA leads for complex enterprise web apps (e‑commerce, configurators): Long, multi‑step flows are brittle and expensive to test manually; they need visual, resilient testing to catch regressions in complex UIs homepage, LinkedIn demos.
• Small startups and product teams with limited QA headcount: They need quick setup, scheduled checks, and low maintenance so QA doesn’t slow shipping; they want recording, scheduling, and preview targeting without heavy scripting Running a Test, CI Integration, homepage.

How would they acquire their first 10, 50, and 100 customers

• First 10: Run founder‑led, concierge pilots with YC peers and early inbound from demo posts; book via the site’s demo flow, set up tests for them, and capture artifacts/time‑to‑fix results to iterate fast YC, LinkedIn, homepage, Quickstart.
• First 50: Offer a limited self‑serve trial with templates and GitHub/CI examples to reduce setup, promoted via targeted community posts, webinars, and product blog; publish short case studies to speed evaluation CI Integration, Quickstart, product blog.
• First 100: Stand up a lightweight enterprise motion (1–2 AEs + a customer success engineer) with paid pilots, security/FAQ docs, and deployment‑blocking CI examples; pursue channel leads with CI/CD and preview platforms, and list in tool directories CI Integration, homepage, LinkedIn.

What is the rough total addressable market

Top-down context:

The global automation‑testing market was about USD 30.24B in 2023 and is projected to reach ~USD 92.45B by 2030 (CAGR ~17.3%) Grand View Research. Within it, web interfaces account for roughly 30% of spend (2024 estimate) Mordor Intelligence.

Bottom-up calculation:

AI‑enabled testing was ~USD 498.4M in 2023, with test automation comprising ~58% (~USD 289M) and projected ~USD 944M by 2030 (58% of USD 1.627B) Grand View Research. Applying the ~30% web share to the broader automation‑testing figure implies a ~USD 9B web‑automation market in 2023 Grand View Research, Mordor Intelligence.

Assumptions:

• Interface split (web ~30%) is applied to the broader automation‑testing total to estimate web‑focused spend.
• AI‑enabled testing ‘test automation’ subshare (~58%) is stable through the forecast window.
• Reported figures from different firms are comparable despite methodology differences.

Who are some of their notable competitors

• Testim: AI‑assisted functional testing with a recorder and ML‑based stability. More code/engineering‑oriented than Docket’s plain‑English prompts; Docket emphasizes vision‑first interaction Testim, Docket blog.
• Mabl: Codeless, ML‑driven browser testing with CI integration and auto‑maintenance. Overlaps on goals but centers on ML test upkeep vs. a vision‑first, natural‑language agent Mabl.
• Applitools: Visual AI testing (visual diffs/validation) that plugs into existing tests. Strong at catching UI regressions but not a full agent for recording or plain‑English test authoring Applitools.
• Playwright (Microsoft): Open‑source, code‑first browser automation used for E2E tests across browsers. Powerful but requires engineering effort vs. Docket’s no‑code, vision‑first agent Playwright.
• Ghost Inspector: No‑code record‑and‑playback browser testing with CI hooks. Similar on recording/scheduling, but lacks Docket’s vision‑first agent and plain‑language authoring focus Ghost Inspector.