Pangolin

Run GitHub Actions on faster bare-metal runners with lower Ubuntu per-minute pricing

Blacksmith is a drop-in replacement for GitHub-hosted runners that executes Actions on bare-metal gaming CPUs and source-shaped cache infrastructure. Migration requires a one-line YAML change. Features include colocated warm caches, persistent Docker layer caching on NVMe, CI observability with log search, and Firecracker microVM isolation. SOC 2 Type 2 certified, with Ubuntu x64 pricing at $0.004/min and 3,000 free minutes/month.

Trusted runtime environments for AI agents in production infrastructure

Teleport Beams provides cryptographically verified, policy-gated access for AI agents to interact with production infrastructure including servers, Kubernetes clusters, and databases. Launched at KubeCon EU 2026, Beams extends Teleport's zero-trust access platform with agent-specific runtime controls, audit trails, and policy enforcement to ensure AI agents operate within defined boundaries when deployed in production environments.

High-performance S3-compatible object storage built in Rust

RustFS is an open-source distributed object storage system built entirely in Rust, offering 2.3x faster performance than MinIO for small object payloads. It provides full S3 API compatibility, enabling seamless migration from MinIO, Ceph, and AWS S3 with existing SDKs and CLI tools. Released under Apache 2.0 license, it avoids MinIO's restrictive AGPL terms. Features include distributed architecture, erasure coding, WORM compliance, encryption via RustyVault, and a web management console.

Zig-built headless browser engineered for AI agent workloads

Open-source headless browser written in Zig for AI agents, crawling, and automation. Lightpanda omits graphical rendering, keeps DOM and JavaScript execution, exposes CDP for Puppeteer/Playwright/chromedp, and adds Agent, PandaScript, and MCP workflows. Current public benchmarks claim about 9x faster execution and 16x less memory than Chrome.

Heterogeneous CPU-GPU inference and SFT for large MoE models

Open-source framework for running and fine-tuning large Mixture-of-Experts models with heterogeneous CPU-GPU execution, optimized kernels, limited VRAM and SGLang or LLaMA-Factory integrations.

Official Kubernetes and Helm reference stack built on the vLLM inference engine

Official vLLM reference implementation for scaling the existing inference engine on Kubernetes with Helm, request routing, KV-cache offload, autoscaling and Prometheus/Grafana observability.

Distributed inference orchestration above vLLM, SGLang and TensorRT-LLM

Open-source, datacenter-scale orchestration layer that coordinates vLLM, SGLang and TensorRT-LLM across nodes with disaggregated serving, KV-aware routing, multi-tier cache management and automatic scaling.

Open-source GPU control plane for scalable AI model serving

Open-source GPU cluster manager that configures vLLM, SGLang, TensorRT-LLM or custom engines, serves models through compatible APIs, and provisions SSH-accessible GPU instances across on-premises, Kubernetes and cloud environments.

Disaggregated KV cache storage and transfer for LLM serving

Open-source infrastructure for disaggregated LLM serving that pools KV caches across prefill and decode workers, with high-performance transfer, distributed storage and integrations for vLLM and SGLang.

Run and govern MCP servers across desktop, CLI and Kubernetes

Open-source MCP runtime and governance platform that runs servers in isolated containers, curates registries, enforces access policies, and operates gateways across desktop, CLI, and Kubernetes.