Continue the conversation — chat opens pre-seeded with the current signal, caps, and movement.
AI-native test automation platform built on AI since 2017, now positioned as 'the agentic testing platform' — coverage that builds itself, runs itself, and recovers itself. Agentic Tester autonomously explores the application, creates tests from natural language or Jira requirements (10x faster claim), and maintains/recovers them via auto-healing (85% maintenance-reduction claim).
Target user: enterprise QA and engineering teams running SaaS, web, mobile, and API applications who want AI-driven test creation/maintenance and can accept cloud-only/US-data deployment. F500 customers (Workday, JetBlue, Barracuda, Charles Schwab, Chewy, DataRobot, NCR, Stack Overflow) and institutional backing (GV, Vista) provide enterprise credibility.
Differentiator: the cloud MCP server is the developer-loop integration surface — debug, create, and run tests from inside Claude Code / Cursor / Windsurf without leaving the AI workflow (May 27 2026 merge-to-root-cause debug). The standalone local MCP server was deprecated May 18 2026 in favor of the cloud MCP. This deepens the developer-loop story relative to traditional test-automation vendors.
Adoption & Proof Points
- **8 named F500 / enterprise customers:** Workday, JetBlue (90% test-execution-time reduction), Barracuda, Charles Schwab, Chewy, DataRobot, NCR, Stack Overflow
- **Analyst recognition:** Gartner Peer Insights 4.7/5 in the AI-Augmented Software Testing Tools market
- **Institutional backing:** GV, Vista Equity, Amplify, Presidio, CRV (last round Nov 2021 Series C, $40M; total raised $76.1M)
- **Growth signals:** revenue ~$17.9M (2024) up from $13.1M (2023, +37% YoY); headcount ~114 (Apr 2026) up from ~94 (2024)
- **Active product cadence:** ~1 release/week in May 2026; cloud MCP local create/run (May 12), local MCP server deprecated (May 18), merge-to-root-cause debug from AI coding agent (May 27)
- **No independent benchmark validation** of agentic-tester / auto-healing claims (85% maintenance reduction, 94% execution-time reduction, 10x faster generation are vendor-reported)
Risks & Limitations
- **`pricing-opacity` cap:** request-a-quote only on mabl.com/pricing (verified 2026-05-28); third-party estimates (~$499/mo Starter) exist but no official public tiers; enterprise pre-sales comparison friction
- **Cloud-only / US-only data residency:** material constraint for EU and regulated workloads; CI-Runner self-hosted execution does not move the data plane
- **No SCIM in public docs, no HIPAA, no ISO 27001:** governance gap relative to enterprise QA buyers
- **AI model governance absent:** LLM provider undisclosed, no documented training opt-out or model selection
- **Vendor-only benchmarks:** no SWE-bench-equivalent for test automation; 85% maintenance reduction / 94% execution-time reduction are vendor-reported
- **Local MCP server deprecated (May 18 2026):** capability preserved in the cloud MCP, but teams that adopted the local server must migrate
- **Series C was November 2021:** ~4.5 years since last raise; offset by revenue growth, but funding maturity worth monitoring
Capabilities & Integration
**Autonomy (13, lower Agentic):** Agentic Tester autonomously explores the app, plans steps, creates/runs/analyzes/maintains tests; auto-healing, Application Summaries (Jan 2026), 10x natural-language generation. Cloud-MCP additions (local create/run, failure investigation, merge-to-root-cause debug) extend the loop without crossing into Self-Directed.
**Integration (13, upper-lower Team-Aware):** GitHub Actions 2.0 + GitHub App + auto issue creation, Jenkins, Bitbucket, GitLab, CircleCI, Jira/Confluence/X-Ray, and the cloud MCP server bridging Claude Code / Cursor / Windsurf. Local MCP server deprecated May 18 2026 — capability consolidated cloud-side. Gap: no PR-workflow ownership, no bidirectional Slack/Linear.
**Context (8, Project-Aware): ** application-surface awareness (UI/DOM indexing, Application Summaries, semantic test search) plus Jira/Confluence requirement context. No source-code understanding, no codebase indexing — appropriate for a testing platform but a hard limit on this dimension.
**Interface (12, Multi-Platform):** web app primary, Desktop App (macOS/Windows/Linux), CLI (Node.js npm, v2.75.x), REST API, cloud MCP server, headless CI-runner mode. No native IDE plugin — the cloud MCP is the IDE bridge.