Quantstruct’s AI platform uses prompts and fine-tuned models that comprise high quality technical documentation and code samples
By Nileena Sunil
Quantstruct, a Y Combinator-backed startup that uses AI for technical documentation, has been launched on Jan. 25. The startup helps improve technical documentation using AI agents researching, testing, and updating documents.
Quantstruct’s AI agents and workflows eliminate documentation debt and have been publishing 100% of change logs and change management updates for hypergrowth companies like Vapi, and enterprises who build API & SDK products, visual software, and more.
READ: Indian American startup launches AI video platform for sales (January 9, 2025)
Founders Sarthak Srinivas and Newman Hu have previously built developer platforms and search infrastructure for multiple Fortune 500 companies, and now, are on a mission to provide better documentation for humans and AI tools to do their best work.
“Newman and I HATED the technical enterprise documentation for 100s of platforms we worked with at our jobs so much, we are starting a company with our friends at Y Combinator to fix it,” Srinivas took to LinkedIn to share prior to Quanstruct’s launch.
“We want to help you build personalized docs for each user, keep your documentation updated with what your product does (according to what your customers ask), and doesn’t require so much manual work to maintain, to start,” he added.
While Srinivas has worked at growth-stage companies like Moveworks as a Kleiner Perkins product fellow where he built an agent developer platform valued over $2 million, Hu has built search infrastructure at Moveworks and Replit where he managed mission-critical platforms powering Gen AI experiences for Fortune 500 companies. Both are published researchers and patent holders from Georgia Institute of Technology’s Computer Science department and University of California, Berkeley’s Electrical Engineering and Computer Science department, respectively.
The founders have identified the problem of documentation debt, where product and technical documents get out of sync with implementation plans. This is caused by a number of factors, from limited time making active efforts for product and customer education difficult, to teams needing to understand too much source code, support, communication threads & product specs to properly guide customers.
READ: Y Combinator startup launch: Unlock AI-driven user assistance with Layup (December 10, 2024)
Quantstruct helps resolve these issues by integrating with the user’s GitHub repositories and Slack workspace. When code changes are detected, the AI analyzes the changes and automatically generates documentation updates. These updates are then sent for review through GitHub pull requests or Slack notifications.
Quantstruct’s AI platform uses prompts and fine-tuned models that comprise high quality technical documentation and code samples. The generated content goes through multiple validation steps and is presented for human review before being published.