Tusk AI, Inc.
Tusk is an AI-powered platform that automates the generation, execution, and maintenance of unit and integration tests for pull requests using codebase and production context.
At a Glance
- Engineering teams at growth-stage startups
- Enterprise companies with large codebases
- Software developers and QA engineers
AI Tools by Tusk AI, Inc.
(1)
Tusk
AI Tests From Production Traffic
Discussions
No discussions yet
Be the first to start a discussion about Tusk AI, Inc.
Latest News
Does Real API Traffic Make AI-Generated Tests Better? An A/B Evaluation of Tusk Drift
Scaling Parallel Agents with the Context-Verification-Harness Framework
What if Your Production Traffic Was Your Test Suite?
How Promptfoo Scaled Test Coverage Without Slowing Development
Products & Services
An AI agent that generates unit and integration tests for pull requests with codebase context.
Generates realistic test cases from live production traffic and business context.
Command-line interface for running and managing Tusk tests locally.
A specialized bot for backfilling unit tests in existing codebases/monoliths.
Market Position
Tusk differentiates itself by leveraging production traffic (Tusk Drift) to generate realistic tests, whereas traditional AI coding assistants focus on code completion. It emphasizes non-blocking PR checks and autonomous test iteration.
Leadership
Founders
Marcel Tan
Co-Founder & CEO. Previously a Technical Product Manager at 6sense, working on their AI email product. Former Company Commander of 300 officers in the Singapore Police Force. Graduated from UC Berkeley.
Sohil Kshirsagar
Co-Founder & CTO. Previously a Senior Software Engineer at Aspire. Graduated from UC Berkeley with a degree in Electrical Engineering and Computer Science. Built a popular online prop bets game for the Super Bowl.
Executive Team
Marcel Tan
Co-Founder & CEO
Product management background at 6sense and Odoo. UC Berkeley alumnus.
Sohil Kshirsagar
Co-Founder & CTO
Software engineering background at Aspire. UC Berkeley EECS alumnus.
Board of Directors
Founding Story
Marcel Tan and Sohil Kshirsagar met at UC Berkeley and combined their backgrounds in product management (at 6sense) and software engineering (at Aspire) to solve the productivity bottleneck of manual testing. They initially built an AI coding agent for general 'chores' before pivoting to focus specifically on high-quality test generation.
Business Model
Revenue Model
Subscription-based SaaS model (Per-seat pricing for developers).
Pricing Tiers
Generation of unit/API tests for PRs, run tests locally or in CI.
Unlimited test generation, CoverBot, observability, Linear/Jira integrations, 3 synced repos.
Minimum 200 seats, self-hosting, SAML/SSO, white-glove implementation.
Target Markets
- Engineering teams at growth-stage startups
- Enterprise companies with large codebases
- Software developers and QA engineers
- Preventing regressions in fast-moving dev cycles
- Increasing code coverage for monoliths
- Backfilling unit tests for legacy code
- Automating regression testing using production traffic
- DeepLearning.AI
- Hamming
- Promptfoo
- TeamFeePay