Mobile teams shipping frequent builds without room to hire more manual testers are the fit. It swaps scripted regression runs for vision-based checks and keeps humans available for subjective calls. Three products cover bug capture, agent verification and hybrid fan-out, and access starts with a demo request.
What changed
QualGent is a Y Combinator Spring 2025 company in San Francisco that sells an autonomous QA agent for iOS and Android apps. It runs reusable test cases against uploaded builds, returns structured bug reports with video evidence, and exposes the work through a REST API limited to 10 requests per second per key.
Maker: QualGent · Autonomy: semi autonomous · Maturity: GA
About QualGent
QualGent is an autonomous quality assurance agent for iOS and Android apps, built by a San Francisco team that is part of Y Combinator's Spring 2025 batch. Its founders say they led engineering teams at Google across Android XR, Stadia, Quick Share and foldable devices, and that earlier research work at NVIDIA, the University of Maryland and Berkeley convinced them that AI could take over repetitive mobile testing. The YC profile lists a team of about 10 people.
What separates it from a script runner is how a test executes. The vendor says it uses computer vision models to drive the app from what is on screen, so a run does not depend on brittle DOM locators, and its About page describes a multi-agent architecture that plans a verification strategy, executes the flow, observes app state and judges whether the behavior was correct. You write a test case once, upload a build, and the agent works through it on a device chosen from the build's platform.
The product is split into three parts. TrustLoop captures bugs from real app usage and turns them into structured reports, repro steps and candidate regression tests. DevLoop gives coding agents, the kind HokAI tracks such as Claude Code, OpenAI Codex and Jules, a way to verify on a device what they just built. QualGent Enterprise fans each test out to an AI or a human verifier and returns one release-readiness view.
Teams that prefer to automate everything use the REST API, whose documented objects are apps, test cases, categories, missions and bug reports. The documented CI pattern uploads a build, submits a mission, polls the request and fails the pipeline on any verified bug, with recipes for GitHub Actions, CircleCI and Jenkins. Bug reports carry steps, expected and actual results, severity, a video attachment and a link to a GitHub, Jira or Linear ticket.
Cloud device farms and test generators cover nearby ground. TestMu AI pairs a cross-browser and cross-device cloud with AI-generated tests, Kusho generates API and UI test suites, and Sennu AI targets Salesforce QA. QualGent publishes no prices and sends buyers to a demo request, and the changelog shows steady API additions through October 2026, including missions, bug reports and a latest-app endpoint.
Pricing
No prices are published. Buyers book a demo, and API use is gated by plan, so the scope you get is set in the sales conversation.
Key Features
- Vision-based test execution: Drives iOS and Android apps from on-screen content, which the vendor says removes the dependence on brittle DOM locators.
- Missions API: POST /v1/missions/submit requests runs of chosen test cases against a build, and the request moves through received, testing, verifying and complete.
- TrustLoop bug capture: Records how a bug appeared during real app usage and drafts the report plus a follow-up test from it.
- DevLoop for coding agents: Gives AI coding agents device-aware verification so they can test what they build before it reaches QA.
- Reusable, versioned test cases: Test cases accept named variables referenced as {{name}} and jobs can pin a historical version for reproducible reruns.
- Real-SMS device routing: A job can set use_sim to run on a SIM-equipped physical device that receives real inbound text messages for one-time-code logins.
Strengths
- The public REST API covers apps, test cases, categories, missions and bug reports, so a CI job can upload a build, request a run and fail the pipeline on a verified bug.
- Each finding arrives with severity, a screen recording and a link to the matching GitHub, Jira or Linear ticket, so developers rarely need to ask for context.
- Pinning a test case to a historical version keeps reruns comparable across releases, which scripted suites often lose.
- Physical devices with SIM cards can receive real SMS codes, which covers login flows that emulator-only tools skip (added April 2026).
Weaknesses
- No prices are published, so every buyer starts with a demo request before comparing cost.
- The docs say a single request can take hours, so it suits nightly or per-release gates better than a fast pre-merge check.
- The device listing endpoint was removed in October 2026 and the device follows from the build, so you cannot choose a phone model through the API.
- The vendor publishes no benchmark or false-positive rate behind its reliability pitch, and the models it runs are not named.
Frequently Asked Questions
What does QualGent cost?
QualGent publishes no prices. The website sends buyers to a demo request, and its API docs return a 403 stating that an organization's plan does not include API access, so what you get depends on the plan agreed with sales. Email sales@qualgent.ai for a quote, and treat any figure you see elsewhere as unverified.
How much of a test run needs a person?
Once you choose the test cases and the build, a request moves through received, testing, verifying and complete without prompting. The Enterprise product can route a check to a human verifier when judgment about look and feel matters. The docs warn that one request can take hours.
Which tools could you use instead of QualGent?
TestMu AI suits teams that want a cross-browser and cross-device cloud plus AI-generated tests. Kusho suits teams whose main gap is API and UI test generation. Sennu AI is the narrower pick for Salesforce QA.
Is QualGent a replacement for a device cloud like TestMu AI?
Not directly. TestMu AI rents you devices and browsers to run your own suites, while QualGent is an agent that executes and judges mobile test cases for you. Choose QualGent when writing and maintaining mobile regression is the bottleneck, and the cloud when you need broad browser coverage.
How do you connect QualGent to a CI pipeline?
Create an API key in the dashboard and send it in the X-Api-Key header. Upload the build, submit a mission for your test cases, then poll the submission and fail the job on any verified bug. The docs include GitHub Actions, CircleCI and Jenkins recipes and suggest polling every few minutes.
Top Alternatives
- TestMu AI: Pick TestMu AI if you also need a cross-browser cloud and a free tier, QualGent if mobile regression is the whole problem.
- Kusho: Pick Kusho to generate API test suites, QualGent to have an agent exercise the finished mobile app on a device.