Case Study
How Mintlify Advances Agentic Documentation With Daytona

1K+
parallel sandboxes provisioned in minutes with Daytona
<90
millisecond sandbox creation times
3
months of engineering time saved
Mintlify is a leading agentic documentation platform that helps technical teams build docs equally readable by humans and AI agents. Backed by Y Combinator, a16z, and Bain Capital Ventures, it powers documentation for thousands of organizations including Anthropic, Coinbase, and PayPal.
Headquarters
San Francisco, CA
Industry
Software Development
Department
Artificial Intelligence Documentation
Key Features
Sandbox Creation Speed Sandbox Statefulness Long‑Running Sandboxes
Learn how this YC-backed documentation platform partnered with Daytona to unlock secure and elastic sandbox infrastructure that powers critical AI agent workflows.
We needed a way to migrate our core agentic workflows to sandboxes, and Daytona didn’t disappoint. Without their lightning-fast provisioning and continuous support, shipping our 'Docs from GitHub' feature in one quarter simply wouldn't have been possible.

Nick Khami
Head of Engineering at Mintlify
01 -- CHALLENGE
Scaling Isolated Agent Deployments Without Heavy Engineering Lift
Mintlify’s agentic platform was created to automate one of the most significant challenges facing technical teams today: keeping documentation up to date. The platform’s AI agents are meticulously trained to understand each organization’s structure and style preferences, while leveraging connected repos to generate in-depth documentation from even the simplest natural-language prompts.
However, to deploy these documentation workflows at scale, Mintlify needed secure, stateful, and long-running sandboxes where agents could analyze code and generate outputs in complete isolation. “The nature of modern-day agents is changing, and most new LLMs are trained to work with native Unix and Linux tooling,” explains Nick Khami, Head of Engineering at Mintlify. “We knew that to give our agents access to that tooling, migrating to sandboxes would be essential.”
While the need for scalable sandbox infrastructure was immediate, Nick knew that building and orchestrating it in-house would be a massive undertaking for his team. Standing up a secure, concurrent sandbox architecture capable of supporting long-running agents would require months of engineering work. Additionally, maintaining a homegrown solution to handle thousands of concurrent runs was simply not a viable option.
Instead of losing momentum to infrastructure management, Nick sought a dedicated runtime platform that would streamline the provisioning of secure, stateful, and long-running sandboxes for key agentic workloads. Given Mintlify’s high production volume and global customer base, the ideal platform would integrate seamlessly without disrupting existing agentic workflows.
During his evaluation of several vendors, Nick discovered Daytona, whose agent-native infrastructure aligned precisely with his needs.
Building and maintaining sandbox infrastructure in-house would have been an incredibly heavy lift for our engineering team. After realizing Daytona was purpose-built for our use case, solidifying the partnership was a no-brainer.

Nick Khami
Head of Engineering at Mintlify
02 -- SOLUTION
A Managed Runtime Platform that Streamlines the Provisioning of Fast, Secure, and Stateful Sandboxes
Thanks to Daytona’s clear documentation and vast list of native integrations, Mintlify rapidly transitioned from an initial proof of concept to a production-ready system. “Daytona had great guides and integrations with tools like OpenCode that were able to leverage to get our system running very quickly,” Nick shares. “The performance we were also seeing right out of the gate was quite impressive.”
Today, Daytona sits at the core of Mintlify’s flagship workflows, most notably their ‘Docs from GitHub’ feature. Now, whenever a Mintlify user wants to generate production-ready documentation from raw GitHub codebases, all they need to do is change the repository URL from github.com/repo to mintlify.com/repo.
On the backend, Daytona instantly provisions a long-running isolated sandbox, clones both the source repo and target docs repo, then runs a coding agent with bash access to scrape README and metadata for branding and project structure. An orchestrator then launches sub-agents to write Markdown pages in parallel, merges outputs with Mintlify CLI validation for links and references, and delivers an instant, branded docs site ready for user tweaks and publishing.
Mintlify also relies on Daytona for inline agentic documentation. When a user asks the Mintlify agent to document a newly resolved customer issue from a Slack thread, it spins up a sandbox, ingests the context, and automatically opens a pull request with the new documentation. On the first run, the system builds a Daytona Volume containing the necessary documentation files. For subsequent requests, they simply remount that existing volume, drastically reducing latency for the end user.
Daytona’s efficiency proved critical at scale. During the launch of their "Docs from GitHub" feature, Mintlify experienced a massive traffic spike. As demand surged, the team used Daytona to dynamically spin up more than 1,000 concurrent sandboxes. Despite Mintlify briefly hitting their account's rate limits, the Daytona team responded within 10 minutes, instantly increasing the quota so the system continued to function under peak load.
While Daytona abstracts away the heavy lifting of infrastructure orchestration, Mintlify retains granular control over all agentic workflows. Using Daytona’s SSH access, Nick and his team can securely tunnel into any active container during development. This low-level visibility enables them to observe exactly how the coding harnesses commands, test hypotheses, and continuously engineer smarter, more reliable workflows.
For any company building agentic workflows in sandboxes, look no further than Daytona. It’s extremely smooth and reliable even when handling thousands of requests.

Nick Khami
Head of Engineering at Mintlify
03 -- RESULT
Mintlify Now Provisions Thousands of Parallel Sandboxes in Minutes With Daytona
With Daytona, Mintlify transformed a complex infrastructure hurdle into a significant strategic advantage. By bypassing months of manual engineering development, the team reclaimed significant bandwidth to focus on continuous platform expansion.
1K+ parallel sandboxes provisioned in minutes with Daytona
<90 millisecond sandbox creation times
[X] months of engineering time saved
Looking ahead, Nick knows Daytona’s sandbox infrastructure will serve as a critical foundation for increasingly larger and more complex agentic deployments.
Our mission is to put documentation on autopilot so technical teams can focus on their core competencies. Daytona makes that vision a reality, empowering us to run smarter, more complex workflows at scale.

Nick Khami
Head of Engineering at Mintlify



