Vcode

English  ·  简体中文  ·  Guide  ·  Spec

A DeepSeek-native AI coding agent for your terminal.

A config- and plugin-driven harness — a single static Go binary, tuned around DeepSeek's prefix cache so token costs stay low across long sessions.

Features

• Config-driven. Providers, the agent, enabled tools, and plugins are all declared in Vcode.toml. No hardcoded models.
• Multi-model & composable. DeepSeek ships as a preset; any OpenAI-compatible endpoint is a config entry, not new code. Optionally run two models together (executor + planner) in separate, cache-stable sessions.
• Plugin-driven. External tools run as subprocesses over stdio JSON-RPC (MCP-compatible). Built-in tools self-register at compile time.
• Cache-aware context maintenance. Startup injects a small stable environment summary, stale tool output is snipped/pruned before summary compaction, and the built-in tool schema contract is documented for regression review.
• Zero-friction distribution. CGO_ENABLED=0 single binary; cross-compile to six targets with one command. The only dependency is a TOML parser.

Install

npm i -g Vcode                  # any OS; pulls the prebuilt native binary
brew install esengine/Vcode/Vcode   # macOS

Prebuilt archives (darwin|linux|windows × amd64|arm64) and SHA256SUMS are on every GitHub release.

Code signing

Windows builds are code-signed with a free certificate provided by the SignPath Foundation, with signing through SignPath.io.

Build from source

make build      # -> bin/Vcode(.exe)
make cross      # -> dist/ (darwin|linux|windows × amd64|arm64)

Quick start

Vcode setup                      # config wizard → ./Vcode.toml
export DEEPSEEK_API_KEY=sk-...      # or let setup save it to Vcode home .env
Vcode                            # then run /init to generate AGENTS.md (project memory)
Vcode run "implement the TODOs in main.go"
Vcode run --model deepseek-pro "add unit tests for this function"
echo "explain this code" | Vcode run

Configuration

A minimal Vcode.toml — one provider and a default model — is enough to start:

default_model = "deepseek-flash"

[[providers]]
name        = "deepseek-flash"
kind        = "openai"
base_url    = "https://api.deepseek.com"
model       = "deepseek-v4-flash"
api_key_env = "DEEPSEEK_API_KEY"

Resolution order is flag > ./Vcode.toml > the user config file > built-in defaults; starting with Vcode v1.8.1, the user file lives at ~/.Vcode/config.toml on macOS/Linux and %AppData%\Vcode\config.toml on Windows. See Configuration paths for migration details and the full config.toml / .env structure. Provider entries name secrets with api_key_env; the secret values themselves live in Vcode's global <Vcode home>/.env, shared by CLI and desktop. Permissions, the sandbox, plugins (MCP), slash commands, @ references, and two-model setup are all in the Guide.

Documentation

• Guide — configuration, permissions & sandbox, plugins (MCP), slash commands, @ references, two-model collaboration.
• Bot guide — connect Feishu, Lark, and WeChat bots from the desktop app, then use approvals, YOLO, and commands from IM.
• Spec — engineering contract: architecture, registries, data types, and roadmap.
• Tool contract — provider-visible built-in tool names, read-only flags, and schema snapshot guard.
• Checkpoints & rewind — the snapshot-based edit safety net (Esc-Esc / /rewind).

---

About This Fork

This fork is based on the upstream project esengine/Vcode with the following modifications:

• All sandbox and permission restrictions removed — the agent can execute commands freely without being blocked by OS-level sandboxing, permission gates, or guardian review. This enables more thorough and efficient task execution.
• Upgraded and optimized on top of the original codebase for a smoother, more convenient user experience.

License

MIT — see LICENSE