Propolis

What do they actually do

Propolis runs swarms of autonomous browser agents that explore a customer’s web app like real users, surface issues, and produce runnable end‑to‑end tests you can keep in CI. It focuses on browser‑based apps and uses Playwright under the hood; teams can export tests as Playwright or run them on Propolis’s infrastructure (propolis.tech, Launch HN).

A typical workflow is: sign in, provide a starting URL and any test credentials, launch a swarm to exercise flows and hunt for errors or flaky behavior, then review issues and accept proposed deterministic tests for ongoing protection in CI (Launch HN). The product is positioned as production‑ready with a quick initial run experience, and they offer exportable artifacts to fit into existing pipelines (Launch HN).

Today it’s focused on web/browser flows (not native mobile). Some flows that depend on receiving emails or third‑party hops may require extra setup, which the team has flagged as an active roadmap area. Early pricing shared publicly is $1,000/month for early customers, with a free initial run to try it (Launch HN).

Who are their target customer(s)

• Small engineering‑led startups shipping web changes frequently without a dedicated QA team: Manual checks and ad‑hoc testing slow releases and miss regressions; maintaining scripted tests consumes engineers’ time (Launch HN).
• Mid‑market product teams with customer‑facing web flows (signups, payments, dashboards): Intermittent or hard‑to‑reproduce bugs in complex user paths slip to production because exploratory testing can’t cover every flow (propolis.tech).
• QA leads responsible for end‑to‑end test stability across many releases: Large suites are brittle/flaky and require constant upkeep, yet still provide limited confidence in real user behavior (Launch HN).
• Platform/DevOps engineers who own release safety and pre‑prod validation: Lack of realistic pre‑production traffic and no cheap, repeatable way to exercise many user journeys as a canary step (Launch HN).
• Heads of engineering with constrained QA budgets: QA headcount and ongoing maintenance of scripted tests reduce product velocity and raise operating costs (propolis.tech).

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

• First 10: Founder‑led outreach to YC networks and Launch HN leads, offering a short trial on the $1,000/mo early plan with hands‑on setup so value is shown in one or two runs; convert wins into 1–2 case studies and quotes.
• First 50: Targeted developer channels (HN, LinkedIn) plus technical content (Playwright export, CI guides), webinars/office hours, and a sales engineer to onboard; publish several detailed customer stories and a simple referral program.
• First 100: Ship self‑serve templates and CI integrations, add usage‑based tiers, and build partnerships (CI/CD platforms, QA consultancies); support with SDR outreach and targeted ads, and double down on the best‑converting channels via instrumentation.

What is the rough total addressable market

Top-down context:

Multiple market studies put software/app testing and automation in the tens of billions; applying an estimated ~30% share for web/browser testing yields a 2024 TAM of roughly $10–$17B for the segment Propolis targets (Business Research Company, Mordor Intelligence, ResearchNester).

Bottom-up calculation:

Using Propolis’s public early price (~$12k/year), the $10–$17B web testing TAM implies a theoretical range of about 0.8M–1.4M potential equivalents if every dollar were tool spend of this type—an upper bound, not a forecast (Launch HN, sources above).

Assumptions:

• Web/browser testing is ~30% of broader software testing/automation spend in 2024 (per market breakdowns).
• Use of 2024 market size figures from cited reports and midpoints where ranges exist.
• $12k/year reflects Propolis’s early public pricing and is used as a simple per‑customer proxy.

Who are some of their notable competitors

• Mabl: AI‑driven testing platform that generates and maintains end‑to‑end tests across web, mobile, and APIs; positioned as a fuller‑stack enterprise platform rather than swarm‑based exploratory agents.
• Testim: AI/ML‑assisted, low‑code UI test creation with an emphasis on stability and TestOps for reliable CI runs, not large‑scale autonomous exploration.
• Autify: No‑code/Playwright‑based platform that uses AI to create and maintain web (and mobile) test scenarios; a direct alternative for automated test generation without writing Playwright by hand.
• Applitools: Visual AI for regression and functional checks with autonomous analysis; often complements or replaces parts of functional testing but isn’t focused on swarm‑style exploratory agents.
• Checkly: Developer‑friendly synthetic monitoring using Playwright browser checks, aimed at continuous scripted checks in production rather than autonomous exploratory sessions.