We treat the line as a hypothesis and the tool call as a measurement. Reasoning is not decoration for action — it is action, made visible at the speed of thought.
LV Agent is built around a Harness microkernel inspired by the Turing machine. It runs multi-round LLM calls inside a tool loop with self-correction: simple questions take few rounds, hard questions take many. The sketchbook is a research instrument — and so is the terminal. A sketch that can be read is worth more than a render that can only be admired.
Multi-strategy reasoning — CoT / ReAct / Verification, with MCTS planning. Adaptive loop control: few rounds for easy asks, many for hard ones.
Web search with multi-query fusion, full file system access, Python & Bash execution with timeout isolation, plus GitHub, PDF, weather, fetch.
Holographic memory: a knowledge graph for durable facts, vector experience memory across sessions, and learned skills extracted from conversation.
Event-sourced execution that is traceable and replayable. SQLite session persistence, dual token/time budgets, checkpoint resume, and hot-swappable modules.
Multi-query fusion across sources, body extraction, scoring, and source evaluation for deep research.
Read / write / grep / glob with cross-directory browsing — “the XXX on my desktop” just works.
Python & Bash with full terminal access — pipes, redirect, && / ||, env — under timeout isolation.
GitHub search, PDF reading, weather query, web fetching, and a standalone Telegram bot.
/deep_researchMulti-angle search → HTML report!cmdExecute shell directly@fileInject file contents into input/modelSwitch model at runtime/strategySwitch reasoning strategy/compressManually compress context/learnLearn a memory skill from chat/memskillManage skills: list / evolve / snapshot/sessionsBrowse historical sessions/dashboardOpen the agent state panel/draftsView stashed input drafts
Prefer the terminal? git clone https://github.com/Xinchen1/LV-Agent.git then ./install.sh for the global lv command.
Backends: DeepSeek · OpenAI · Anthropic · OpenRouter · Ollama (offline).
LV Agent runs in the browser — no install required. Bind a local folder, generate PDFs, write code, search the web. Your files stay on your disk.
Select any folder on your computer. The agent reads, writes, and creates files directly on your disk via the File System Access API.
Ask the agent to create a report, and it renders a formatted PDF with Chinese font support — saved straight to your chosen folder.
Web search, code execution, API calls, browser automation — the same 16+ tools as the terminal version, running in a container.
Watch the agent think in real-time. Generated files appear in the Artifacts panel — click to open or download.
brew install --cask https://github.com/Xinchen1/LV-Agent/releases/download/v1.0.0/LV.Agent-1.0.0-arm64.dmg
# or clone & install
git clone https://github.com/Xinchen1/LV-Agent.git
cd LV-Agent && ./install.sh
Gives you the global lv command.
export LV_API_KEY="your-key-here"
# Or edit config.yaml
vim config.yaml
Backends: DeepSeek · OpenAI · Anthropic · OpenRouter · Ollama.
lv
# Start chatting
> 搜索 Python 3.13 的新特性并生成 PDF 报告
# Or use the web edition
open https://lv-agent-web.onrender.com/
The agent picks tools automatically based on your task.
Use /deep_research for multi-angle research with a synthesized HTML report. Use !cmd to run shell commands directly.
No install needed — just open the web edition, click 📁 to bind a folder, and start working.
“ The hand that draws and the mind that reasons are the same instrument — sharpened from different ends. — studio note, plate VII