English | 中文
LLM Space 4
ScreenFlow.mp4
LLM Space v4 is a desktop app for agent builders — prototype your next agent ideas, inspect every step of your harness execution, debug failures, and evaluate performance, all in one place.
Official website: https://deer-flow.github.io/llm-space/
LLM Space is a sister project of DeerFlow, and we dogfood it heavily: every version of DeerFlow is built and debugged with LLM Space. The project started in March 2023, and v4 is its fourth major iteration.
Contents
- Features
- Tech stack
- Project layout
- Download
- Install
- Run the app
- User guide
- Contributing
- Sponsors
- Donate
- License
Features
- Build — write and version your prompts, system messages, tools, and model settings.
- Trace — see every model call and tool run inside the agent loop as it happens.
- Debug — replay a run from history and step through it to find what went wrong.
- Evaluate — measure how your agent performs across runs.
- Manage — keep your threads organized as files on your own machine.
- Generate — let AI write your prompt and tools for you, and even turn any thread into a runnable LangGraph agent.
Your files and API keys stay on your local computer. LLM Space collects a small amount of anonymous usage data to improve the app - see TELEMETRY.md for exactly what is collected and how to opt out.
Tech stack
- Language & tooling — TypeScript, built and managed with Bun.
- Desktop shell — Electrobun, a lightweight way to ship a native app.
- UI — React with Tailwind CSS and shadcn/ui.
- Agent framework — Pi Agent Core, a lightweight agent framework for building agents.
Project layout
LLM Space is a Bun monorepo:
packages/
core/ # Shared domain types, clients, storage, and generators
runtime/ # Local runtime, models, tools, skills, MCP, and Plugins
ui/ # Shared React design system and Thread Playground
apps/
desktop/ # The desktop app (Electrobun shell + React UI)
examples/
atlas-plugin/ # Complete Plugin example covering every Extension type
Each package keeps tests in a tests/ directory beside src/. Test paths
mirror source paths, such as packages/core/src/thread/history.ts and
packages/core/tests/thread/history.test.ts.
Download
Grab a DMG from the latest release — macOS, Apple Silicon and Intel. It comes in two editions:
- LLM Space — uses the system WebView. Small download (~27 MB), light on memory and battery.
- LLM Space Performance — embeds its own rendering engine (~130 MB). Rendering stays consistent across macOS versions, and usually performs better.
Install either, or both. They share the same ~/.llm-space data, so switching editions keeps your threads and settings, and both update themselves in place.
Install
Building from source? You need Bun first. Bun is a fast, all-in-one runtime and package manager for JavaScript — think of it as a drop-in replacement for Node.js and npm. Follow the official install guide.
Once Bun is ready, install the project from the repo root:
bun installContributing, or want the exact toolchain CI uses? Install mise and run mise run setup instead — it installs the locked Bun version (from mise.lock) plus JS deps in one step.
Run the app
Start the desktop app for local development:
mise run devBuild a canary release:
mise run build:canaryUser guide
The user guide lives in this repository:
- Quick start
- User manual
- Core concepts
- Conversation compaction
- Sharing Threads
- Generating Projects
- Plugin development guide
- Complete Atlas Plugin example — two Skills, MCP servers, model providers, Plugin Tools, Commands, and Thread Storages, plus a multi-field Settings schema.
Contributing
For now, we only merge pull requests from the DeerFlow core team members.
Everyone else is very welcome to help by opening an issue — bug reports, ideas, and feedback all make the project better.
Sponsors
LLM Space is free and open source, and it stays that way thanks to our sponsors. We are proud and grateful to be backed by them.
🏆 Platinum sponsor
Superdesign is an AI product design agent that turns natural-language prompts into UI mockups, components, and full designs on an infinite canvas. Thank you for making LLM Space possible. 💜
Want to see your logo here? We would love to talk - reach out by opening an issue or support the project.
🏅 Gold sponsor
We strongly recommend using the VolcEngine Coding Plan as your default model provider:
Donate
If LLM Space is useful to you and you would like to support its development, you can donate here:
Thank you.
License
LLM Space is released under the MIT License.
