Docs tutorial on running browser agents inside Runloop devboxes using Browserbase-managed browsers connected via Playwright's CDP protocol, with no local Chromium install required. Covers multi-page research with session reuse and AI-powered automation via Stagehand.

Docs tutorial on running browser agents inside Runloop devboxes using Kernel-managed browsers, with no local Chromium install required. Covers multi-page research within a single browser session and computer-vision-based agent control (Claude and OpenAI computer-use models).

Walkthrough of Runloop's Benchmark Orchestrator and RLI (Runloop's CLI) for running AI benchmarks at scale: launching parallel benchmark jobs across multiple models, monitoring runs in real time, and reviewing scored results in the dashboard.

Demo of Runloop's integration with OpenAI's Agents SDK using a tax-prep workflow: isolated, resumable devboxes handling document uploads and multi-step agent reasoning inside secure sandboxes.

Announced Runloop's participation in Stripe Projects' developer preview, a co-designed "Agentic Provisioning Protocol" that lets developers provision a full Runloop account (devboxes, orchestrator, credentials) with a single CLI command.

Terminal walkthrough provisioning Runloop's Pro Plan directly through the Stripe Projects CLI: from installing the CLI to confirming the link and switching over to the Runloop CLI to use the newly provisioned environment.

Final entry in the AI TaxMan series: building custom, reproducible benchmarks to evaluate and improve agent performance, with parallel execution on Runloop devboxes and Weights & Biases Weave tracing.

Second entry in the AI TaxMan series: experimenting with six prompt variations in identical Runloop devbox environments and using Weights & Biases Weave for automated tracing and observability.

First entry in the AI TaxMan series: an open-source agent that parses W2s and generates Form 1040s inside isolated Runloop devboxes, built to show why secure, auditable execution environments matter for agents handling sensitive data.

Tutorial on bootstrapping LangChain Deep Agents from a local prototype to a deployed, shareable production app in under 10 minutes, using Runloop devboxes for durable, reliable infrastructure.

Multi-agent content-production system: orchestrated LangChain/LangGraph entities automating a multimedia/video content pipeline end-to-end. Built and demoed as a technical preview; no longer running live.

Technical preview of a quant-analysis assistant applying LangChain/LangGraph agent orchestration to investment research and trading workflows.

Agentic memory manager for AI agents, with configurable, compartmentalized memory banks for RAG-style storage and retrieval across sources.

Technical preview of VIA, a job-search automation pipeline that syncs job boards and drives an agent-assisted application/career-transition workflow.

A web platform exploring common complex AI scenarios as simple, usable products, the product arm of the Tesseract Agentic & Automation Studio, also home to Tesseract Academy.

An AI-agent-powered job search platform and companion Chrome extension that automates finding, applying to, and tracking job opportunities end to end.

A personal, later open-sourced dev tool exploring modern RAG and long-term memory concepts: a consistent ingestion and retrieval experience across files, documents, archives, and internal knowledge bases.

A library of 300+ production-ready AI automation templates for engineers and developers, systemized from client automation work into a self-serve product covering n8n, Flowise, and LangChain workflows.