Hoplite (YC S26) Launches Cloud Deployment for Coding Agents with QA Previews
Hoplite (hoplite.sh) is a new service from YC S26 that deploys coding agents in the cloud with a focus on easily QA-ing features. It ports your local setup—sessions, memories, MCP servers—and gets your projects ready to run in the cloud. The founders, Bence and Ryan, built a custom agent harness instead of using off-the-shelf solutions like Codex or Claude Code, to have independence and test new features without waiting for Anthropic or OpenAI.
Key Features
- Custom agent harness: Built in-house, giving full control and flexibility to experiment with new agent behavior.
- Cloud infrastructure: Hosted on AWS, with Temporal for durable workflows, Modal for sandboxes, and Planetscale for the database. They emphasize reliability and security as agents become tier-0 infrastructure.
- Onboarding: Transfers your local sessions, memories, and MCP servers to the cloud, making setup seamless.
- Previews for QA: The team is optimizing previews to let developers visually verify new user flows, UI, API responses, and CLI behavior across platforms like Windows, while running hundreds of agents concurrently.
Pricing and Trial
You can try Hoplite for free with the code HACKERNEWS, which includes $100 in free credits. You can also connect your Codex subscription to use OpenAI models through Hoplite. More details at hoplite.sh/pricing.
Who It's For
Developers who want to run AI coding agents at scale in the cloud and need an efficient way to QA the results, especially those who expect to review product output rather than code as models improve.
📖 Read the full source: HN LLM Tools
👀 See Also

Tessera: Open-Source GUI Workspace for Managing Multiple Claude Code Sessions
Tessera is an open-source GUI that lets you run multiple Claude Code sessions side by side with Git worktree isolation, Kanban task tracking, live diffs, and agent activity inspection.

Claude Code v2.1.141: New Environment Variables, Hooks Enhancement, and Bug Fixes
Anthropic released Claude Code v2.1.141 with new environment variables (CLAUDE_CODE_PLUGIN_PREFER_HTTPS, ANTHROPIC_WORKSPACE_ID), terminalSequence field for hooks, agent listing by cwd, and over 20 bug fixes.

Open Source AI Agent Prompt Library Reaches 100 GitHub Stars
A community repository called ai-setup provides shared system prompts, Cursor rules, Claude configs, and local model workflow setups for AI agents. The project has 100 GitHub stars and 90 merged PRs.

Routerly: Self-Hosted LLM Gateway with Runtime Routing Policies and Budget Control
Routerly is a free, open-source, self-hosted LLM gateway that provides runtime model selection based on routing policies like cheapest, fastest, or most capable, along with project-level budget limits with per-token tracking. It's OpenAI-compatible for drop-in use with tools like Cursor, LangChain, and Open WebUI.