Playgent

We build environments to test your app and agents against the real world. Deploy your code/agent into a sandbox that runs against replicas of external services (Stripe, Slack, Google Drive, etc.). Twins support the same APIs, MCP tool calls & SDKs as the real software, and you can spin up thousands of instances in parallel and seed them using natural language. Run real workflows, surface edge cases, and observe how your code/agents actually behave before hitting production.

shared product terms: environment, test, sandbox, mcp; shared product theme: developer infrastructure; same category: AI agents; shared affiliation: Y Combinator

Physical Turing tests humanoids in the real world and returns actionable failure data before deployment. We procure, staff, and operate a wide range of target environments for real-world rollouts, helping robot makers, model companies, and enterprises continuously evaluate where humanoids succeed, where they fail, and what needs to improve before each deployment.

shared extracted concept terms: enterprise; shared product terms: test, failure, environment, enterprise; shared affiliation: Y Combinator

Boundary is building BAML -- a programming language to build AI agents. We used to code in assembly, and moved on to C. From C, we moved on to higher level languages like Python. Now we are going from Python to natural language. What does the code look like when 50% of the business decisions is decided by an AI agent / prompts? How do you test these AI components? We made BAML to address these problems. BAML helps has built-in tools to test, observe, and work with structured LLM outputs / tool c

shared product terms: language, natural, what, when; shared product theme: developer infrastructure; same category: AI agents; shared affiliation: Y Combinator

Plotdrive is the first continuous production system for narrative — infrastructure that transforms how long-form stories move from imagination to completion. Most books don’t fail for lack of talent; they fail when the workflow collapses somewhere between page 80 and page 300. Founded by brothers with backgrounds in high-stakes poker, VR, and narrative systems, we built what we couldn’t find: an engine that holds the entire shape of a story while authors work inside it. Writers see 5-10× gains i

shared product terms: when, fail, production, what; shared product theme: developer infrastructure; shared affiliation: Y Combinator

Understand user patterns and find silent failures in your AI agents. The Context Company helps teams analyze AI agent conversations to surface user behavior patterns, performance trends, and moments where agents struggle - so teams know what’s working, what’s not, and what to improve next. We turn raw conversations into structured signals: topic clustering, user feedback analysis, custom pattern tracking, and alerts. All alongside traditional observability like traces, tool calls, latency, and c

shared extracted concepts: tool calls; shared product terms: failure, call, context, what; shared product theme: context and memory; same category: AI agents; shared affiliation: Y Combinator

Narrative is the easiest way to create end-to-end tests, replacing Playwright, Selenium, and Cypress with self-healing tests. Record your user journeys to deploy tests in minutes, and let our agents handle flaky selectors and timing issues. When a test fails, you get a bug report in plain English that's easy to triage. Trusted by Nanonets, Commenda, Drumkit, Metal, and Vega. Learn more at www.trynarrative.com

shared product terms: create, test, when, fail; shared product theme: context and memory, developer infrastructure; same category: AI agents; shared affiliation: Y Combinator