Legal
Service Description
What the Minitap platform includes: its services and features, support and availability commitments, how data is handled, and what is out of scope.
Last updated: October 5, 2026
1. Overview
Minitap provides mini, an AI QA agent that writes, runs, and maintains end-to-end tests for iOS, Android, and web apps. mini takes on repetitive testing work such as regression packs, test upkeep, and bug reproduction. Customer teams stay in charge of scenarios, review, and the ship decision.
This Service Description covers the paid Minitap platform: the web dashboard at app.minitap.ai, the command-line tool (CLI), and the CI integrations. It does not cover the open-source mobile-use SDK, which is published separately under the Apache 2.0 license.
The contracting entity is Minitap, Inc. (Delaware), unless the order form names MINITAP SAS (France) or MINITAP LIMITED (UK). The customer's order form, the Terms of Use, any service level agreement, and any data processing agreement (DPA) take precedence over this Service Description if they conflict.
2. Services and features
The platform has three services, all included on every plan.
2.1 mini, the AI QA agent
- Generates test suites from product specs, the app, and the repository
- Runs tests on pull requests and release candidates, and on demand from CI, the CLI, or Slack
- Maintains tests as the app changes, and flags any change a human must approve
- Produces reports with a verdict, screenshots, video recordings, and a trace of what mini did on each run
- Triages failures and reproduces bugs
- Suggests UI and UX improvements and checks accessibility against WCAG 2.2
2.2 Test execution
| Target | Environments | Accepted builds |
|---|---|---|
| iOS | Simulators, managed physical devices, a shared device pool, or the customer's device lab | Simulator .app or .ipa |
| Android | Emulators, managed physical devices, a shared device pool, or the customer's device lab | .apk with x86-64 libraries, or .apks |
| Web | Chrome, Firefox, and Safari at mobile, tablet, and desktop viewports | Staging or preview URL |
Supported app frameworks are Swift, Kotlin, React Native, and Flutter. A scenario can drive up to 3 devices. Scenarios can also cover GPS location, offline mode, deep links, push notifications on iOS simulators, saved test accounts, and email one-time passwords or magic links received in mini's own QA inboxes. mini can also build the app from the customer's repository, so no build has to be uploaded.
2.3 Access and integrations
- Web dashboard at app.minitap.ai
- CLI, installable with Homebrew, uv, or curl, to manage apps, specs, builds, and runs
- GitHub Action that signs in with GitHub OIDC, uploads or builds the app, waits for a verdict, and reports through Check Runs
- CI systems: a native GitHub Action; any other CI system, such as GitLab CI, Bitbucket Pipelines, Azure Pipelines, CircleCI, or Jenkins, through the CLI
- Source control: GitHub (pull request checks and comments); GitLab and Bitbucket (repository connection and builds)
- Specs and tickets: Notion; Jira and Linear on request
- Collaboration: Slack (start runs by mentioning mini, and receive results) and Figma
- Agents: an MCP server that coding agents such as Claude Code and Cursor can use to drive mini
3. Plans and deployment options
Plans, capacity, and included features are listed on the pricing page. What applies to each customer is set in its order form.
Team and Scale run on Minitap's managed cloud, hosted in the EU or the US at the customer's choice. Enterprise customers can also contract a custom deployment:
- A dedicated single-customer environment in an agreed region
- Bring your own cloud (a private VPC on AWS, GCP, or Azure)
- A hybrid setup, with data at rest in the customer's cloud
- On-premises, with or without Kubernetes
- Air-gapped environments
For custom deployments, scope, timeline, and responsibilities are set in a separate statement of work.
4. Support, availability, and maintenance
Customers get support in a dedicated Slack channel. Live service status and incident history are published at status.minitap.ai.
| Topic | Team and Scale | Enterprise |
|---|---|---|
| Support channel | Dedicated Slack channel | Named owner in the customer's Slack |
| Support hours | Monday to Friday, 9:00 to 18:00 CET, excluding French public holidays | Same |
| First response, P1 | 4 business hours | 2 business hours |
| First response, P2 | 1 business day | 4 business hours |
| First response, P3 | 2 business days | 1 business day |
| Monthly uptime | 99.5% target, no service credits | 99.5% commitment, with service credits |
| Documentation | minitap.ai/docs | Same |
4.1 Priority levels
| Priority | Definition | Examples |
|---|---|---|
| P1, critical | The service is unavailable, or runs cannot start or finish, with no workaround | Dashboard down, every run stuck in queue, CI Action failing for all customers |
| P2, major | A main feature is degraded, but a workaround exists or only part of the service is affected | iOS runs fail while Android works, reports are missing videos, a Slack integration is down |
| P3, minor | A question, a cosmetic bug, or a request with no impact on testing | A wrong label in the dashboard, a how-to question, a feature request |
The customer proposes the priority when reporting, and Minitap may adjust it using these definitions.
4.2 Uptime
The core service is up when customers can sign in to the dashboard, call the API, and start and complete test runs. A minute counts as downtime only when the core service is unavailable, which is a P1 situation. A degraded feature (P2 or P3) does not count as downtime. Uptime is measured by the external monitoring behind status.minitap.ai.
Uptime excludes planned maintenance announced at least 48 hours in advance on the status page, outages of third-party providers outside Minitap's control, and problems caused by the customer's builds, environments, or test accounts.
4.3 Service credits
For Enterprise customers, a service credit is a discount on the next invoice when monthly uptime falls below the commitment: 5% of the monthly fee for each 0.5% below 99.5%, capped at 25%. For example, at 99.0% uptime the customer receives 5% off. Service credits are the customer's only financial remedy for downtime.
4.4 Maintenance and product changes
- Minitap ships updates continuously. Most releases need no downtime and no action from the customer.
- Planned maintenance that affects availability is announced on the status page at least 48 hours in advance.
- Minitap emails workspace admins 30 days before any change that restricts the services described here. Improvements and additions need no prior notice and are recorded in the change log below.
5. Security and data handling
The binding details live in one place each. This section summarises them and links to the source.
| Topic | Summary | Details |
|---|---|---|
| Hosting | Team and Scale run on Minitap's managed cloud in the EU or the US, at the customer's choice. Enterprise can choose a custom deployment (section 3). | Trust center |
| AI models | Inference goes through OpenRouter to Anthropic, Google, OpenAI, ByteDance, and Alibaba models. Zero data retention with the model providers can be enabled on request on Scale and Enterprise. | Privacy Policy, section 4 |
| Use of data to improve the service | Aggregated usage metrics are used to run and improve the service. Minitap does not use customer content to train models, on any plan. | Terms of Use, section 6 |
| Data categories, retention, and deletion | Customer content is deleted within 30 days of contract end. | Privacy Policy |
| Security measures and subprocessors | Encryption, access control, and the subprocessor list | DPA and trust center |
| Compliance | SOC 2 Type II audit in progress | Trust center |
Security issues can be reported to security@minitap.ai.
6. Limitations and exclusions
mini supports the customer's own review and does not guarantee that every defect is found. The following are outside the service unless the order form says otherwise:
- Security, penetration, load, and performance testing
- Custom development on the customer's app
- Beta features and trials, which are provided as is, with no service level commitment
6.1 Customer responsibilities
- Provide a staging or preview environment, test accounts, and any test bypasses mini needs
- Never give Minitap production credentials, payment card data, health data (PHI), or children's data
- Review mini's verdicts and suggested changes before acting on them, since AI output can be wrong
- Keep connected integrations authorised, and revoke access when it is no longer needed
- Comply with the Terms of Use
7. Document governance
This Service Description is owned by Minitap's Chief Product Officer. It is reviewed every quarter and whenever plans, features, or hosting change.
| Date | Version | Change |
|---|---|---|
| September 27, 2026 | 1.0 | First published version |
| October 5, 2026 | 1.1 | No model training on customer content; managed cloud in the EU or the US; integrations and run triggers clarified |