Runtime

RunRL improves LLMs and AI agents with reinforcement learning. Do you have a model that needs to get better at a certain task? Are you tired of messing around with prompts? Do you spend a lot of money on observability, and wish that all this data could let your model self-improve? Well, head on over to RunRL.com, and see what we have to offer! If you give us a model, a prompt, and a reward, we’ll make your model’s reward go up.

shared extracted concept terms: observability; shared product terms: observability, let, com, see; same category: AI agents; shared affiliation: Y Combinator

Respan (formerly Keywords AI) gives teams a unified control plane to trace and evaluate agent behavior without guesswork, automatically surface issues, and fix what breaks faster. Respan connects production observability, automated and human-in-the-loop evaluations, and an adaptive AI gateway to close the loop between detection, decision, and action - so agents improve continuously in production. Respan is trusted by 100+ AI startups and enterprise teams. Today, the platform process 1B+ logs and

shared extracted concept terms: observability; shared product terms: give, control, what, break; same category: AI agents; shared affiliation: Y Combinator

If your team uses Cursor, Claude Code, or Codex to ship, you've felt it: the AI is great at writing code, mediocre at debugging it. When something breaks in prod, you're back to grepping logs, adding console.logs, redeploying, and hoping the AI's guess is right. HyperProbe fixes this. Your coding agent dynamically drops non-breaking probes into your running services, straight from your IDE. When a probe fires, HyperProbe captures and sanitises the exact variable state, feeding your agent real ev

shared extracted concept terms: coding; shared product terms: ship, break, prod, re; same category: AI agents; shared affiliation: Y Combinator

In the space of AI coding agents, Anima bridges the LLM design gap. With over 1.5 million installs, Anima is the first option for professional design, product, and R&D teams going from design to code. Thanks to On-brand Vibe-coding, anyone in the team can build fast without breaking your look and feel. No Figma skills required. Anima generates clean and developer-friendly code. So your prototypes could also be used for handoff. Three ways to start - Figma to Code → Instantly generate clean React

shared extracted concept terms: coding; shared product terms: coding, anyone, fast, break; same category: AI agents; shared affiliation: Y Combinator

Alter is a zero-trust identity and access control platform purpose-built for AI agents. It wraps every tool call in strong authentication, fine-grained authorization, and real-time guardrails, so agents can move fast without breaking things. Each request is verified at the parameter level, authorized against granular policies, executed with least-privilege access, and fully audited in real time. Unsafe actions, whether it’s a rogue DROP TABLE or a payment above policy limits, are blocked before

shared extracted concept terms: guardrail; shared product terms: control, guardrail, so, fast; same category: AI agents; shared affiliation: Y Combinator

Ficra enables product teams to ship higher quality releases, faster. Ficra is the source of truth for every product screen and user flow, giving companies real-time visibility into what their users actually see in their live product. Ficra's AI provides a live, visual map of your users’ journeys: every screen, flow, A/B test, push notification and email. With Ficra, product teams no longer waste 25% of their time auditing complex user flows to assess product quality and find the blind spots in f

shared product terms: enable, ship, visibility, what; shared affiliation: Y Combinator