[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-93958":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":9,"language":10,"languages":9,"totalLinesOfCode":9,"stars":11,"forks":12,"watchers":13,"openIssues":14,"contributorsCount":15,"subscribersCount":15,"size":15,"stars1d":15,"stars7d":15,"stars30d":15,"stars90d":15,"forks30d":15,"starsTrendScore":15,"compositeScore":16,"rankGlobal":9,"rankLanguage":9,"license":17,"archived":18,"fork":18,"defaultBranch":19,"hasWiki":18,"hasPages":20,"topics":21,"createdAt":9,"pushedAt":9,"updatedAt":22,"readmeContent":23,"aiSummary":24,"trendingCount":15,"starSnapshotCount":15,"syncStatus":25,"lastSyncTime":26,"discoverSource":27},93958,"AgentENV","kvcache-ai\u002FAgentENV","kvcache-ai","AgentENV (AENV) is a distributed platform for running agent environments at scale.",null,"Rust",2900,231,10,28,0,59.1,"MIT License",false,"main",true,[],"2026-09-21 04:01:26","\u003Cdiv align=\"center\">\n  \u003Cpicture>\n    \u003Csource media=\"(prefers-color-scheme: dark)\" srcset=\"assets\u002Fheading-logo-dark.svg\" \u002F>\n    \u003Cimg src=\"assets\u002Fheading-logo.svg\" alt=\"AgentENV\" \u002F>\n  \u003C\u002Fpicture>\n  \u003Cp>\u003Cstrong>Running agent environments at scale\u003C\u002Fstrong>\u003C\u002Fp>\n\u003C\u002Fdiv>\n\n[![Coverage workflow status](https:\u002F\u002Fgithub.com\u002Fkvcache-ai\u002FAgentENV\u002Factions\u002Fworkflows\u002Fcoverage.yml\u002Fbadge.svg?branch=main&event=push)](https:\u002F\u002Fgithub.com\u002Fkvcache-ai\u002FAgentENV\u002Factions\u002Fworkflows\u002Fcoverage.yml)\n[![Latest coverage report](https:\u002F\u002Fgithub.com\u002Fkvcache-ai\u002FAgentENV\u002Fblob\u002Fcoverage-data\u002Fcoverage\u002Fbadge.svg?raw=1)](https:\u002F\u002Fgithub.com\u002Fkvcache-ai\u002FAgentENV\u002Fblob\u002Fcoverage-data\u002Fcoverage\u002Fcoverage.json)\n📖 [Full documentation](https:\u002F\u002Fkvcache-ai.github.io\u002FAgentENV\u002F)\n\nAgentENV (AENV) is a platform for running agent environments at scale, powering agentic RL training for **Kimi K3**.\n\n---\n\n## 🚀 Why AgentENV\n\n- **Scale across diverse environments**: AENV runs massive numbers of Firecracker environments across machines and diverse OCI-compatible images, loaded on demand via [overlaybd](https:\u002F\u002Fcontainerd.github.io\u002Foverlaybd\u002F#\u002F). Local disk acts as a bounded cache, retaining hot data and evicting cold, so images can exceed disk capacity while startup stays fast cluster-wide, without pre-warming every host.\n- **Make idle environments inexpensive**: Snapshot-backed environments boot or resume in under 50 ms and pause in under 100 ms. Idle environments can quickly release CPU and memory, then return when new work arrives.\n- **Native snapshot and fork support**: AENV snapshots memory and filesystem changes incrementally, completing in under 100 ms even under heavy disk modification. A running environment can fork into multiple independent sandboxes for parallel agent workflows. Snapshots persist to S3-compatible object storage or a shared distributed filesystem to prevent data loss.\n- **Preserve performance and density over time**: AENV delivers high-performance I\u002FO via ublk while sharing the host page cache across storage and memory-snapshot data. Memory ballooning returns reclaimable guest memory to the host, sustaining high overcommit as environments run longer and diverge.\n\n---\n\n## 📋 Prerequisites\n\n- **Linux kernel 6.8+**; the install script additionally requires **Ubuntu 24.04** (see *Quick Start* below for installation options)\n- `\u002Fdev\u002Fkvm` access for Firecracker microVM execution\n\n---\n\n## ⚡ Quick Start (Single Node)\n\n> [!WARNING]\n> **AgentENV currently does not support authorization.** Do not expose the AgentENV\n> API to the public network. Run it only on a trusted network or behind an\n> authorization proxy with appropriate network controls.\n\n**1. Install and start the server**\n\n*Option A — install script (Ubuntu 24.04)*\n\nInstall both the server and the `aenv` CLI, then start the server as a systemd service:\n\n```bash\ncurl -fsSL https:\u002F\u002Fraw.githubusercontent.com\u002Fkvcache-ai\u002FAgentENV\u002Fmain\u002Fscripts\u002Finstall.sh | sudo bash\nsudo systemctl start aenv\n```\n\n*Option B — Docker*\n\nSet up the server:\n\n```bash\ncurl -fsSL https:\u002F\u002Fraw.githubusercontent.com\u002Fkvcache-ai\u002FAgentENV\u002Fmain\u002Fscripts\u002Fdocker-setup.sh | sudo bash\ndocker pull ghcr.io\u002Fkvcache-ai\u002Faenv-server:latest\ndocker run -d --privileged -v \u002Fdev:\u002Fdev -p 8000:8000 ghcr.io\u002Fkvcache-ai\u002Faenv-server:latest\n```\n\nThe server is accessible at `http:\u002F\u002F127.0.0.1:8000` by default.\n\n**2. Install the aenv CLI** *(skip if you used Option A in step 1)*\n\nInstall separately if you used the Docker method above, or if you are running\nthe CLI on a different machine from the server. Supports Linux and macOS on\nx86_64 and arm64:\n\n```bash\ncurl -fsSL https:\u002F\u002Fraw.githubusercontent.com\u002Fkvcache-ai\u002FAgentENV\u002Fmain\u002Fscripts\u002Finstall-cli.sh | bash\n```\n\n**3. Authenticate**\n\n```bash\naenv auth\n# AENV server URL [http:\u002F\u002Flocalhost:8000]: http:\u002F\u002F127.0.0.1:8000\n# API key: dummy\n```\n\n**4. Pull a template and run a sandbox**\n\n```bash\naenv pull ubuntu:22.04 --name ubuntu\naenv start ubuntu            # starts a sandbox and attaches an interactive shell\n```\n\n---\n\n## 🗂 Deployment\n\nFor Docker Compose \u002F Kubernetes cluster deployment and build-from-source instructions,\nsee 📖 [Deployment](https:\u002F\u002Fkvcache-ai.github.io\u002FAgentENV\u002Fdeployment\u002Fmanual-compile.html).\n\n---\n\n## 🔌 E2B compatibility\n\nAgentENV exposes an E2B-compatible HTTP API. Point `E2B_API_URL` at your\nserver and use the standard E2B Python \u002F TypeScript SDK without any code\nchanges. See 📖 [E2B integration](https:\u002F\u002Fkvcache-ai.github.io\u002FAgentENV\u002Fintegration\u002Fe2b.html)\nfor setup details.\n\n---\n\n## 🛠 aenv CLI reference\n\n```bash\n# Templates\naenv pull docker.io\u002Flibrary\u002Fubuntu:latest --name ubuntu    # FROM \u003Cimage> → template\naenv template list                      # alias: aenv template ls\n\n# Sandboxes\naenv start ubuntu                       # start + attach interactive shell\naenv start ubuntu --detach              # start, print sandbox ID, don't attach\naenv cn \u003Csandbox-id>                    # reattach a shell\naenv exec \u003Csandbox-id> ls -la \u002F         # one-shot command\naenv ls\n\naenv pause   \u003Csandbox-id>\naenv resume  \u003Csandbox-id>\naenv timeout \u003Csandbox-id> 600           # extend TTL to 600 s from now\naenv delete  \u003Csandbox-id>               # alias: aenv rm\n```\n\n`aenv start` accepts a template UUID or human-readable name\u002Falias. `aenv list`\noutputs a table on TTY and JSON when piped; override with `--output table|json`.\n","AgentENV 是一个面向大规模智能体（agent）环境运行的分布式平台，专为高密度、低延迟的智能体强化学习训练场景设计。它基于 Firecracker 微虚拟机与 overlaybd 按需镜像加载技术，支持跨节点部署 OCI 兼容环境；提供毫秒级快照\u002F恢复（\u003C100ms）、内存与文件系统增量快照、快速 fork 沙箱、以及基于 ublk 的高性能 I\u002FO 和内存气球回收机制。适用于需要并行运行数千个隔离、可复现、资源可控的智能体沙箱环境的场景，如大模型智能体 RL 训练、自动化评测与仿真测试。",2,"2026-08-06 03:55:40","CREATED_QUERY"]