Viden CLI
Viden is a terminal cockpit for a coding agent: chat with a model, approve edits and supervise Codex and Claude in one window.
- 86 ★ on the repository
- Last commit
Broad access
Why this level
- The agent runs shell commands and writes files in the workspace, and does so without prompts when approvals are off.
- Delegated external agents are also launched from the cockpit.
Install
- Works in
Manual install
brew install wikieden/tap/videnRecommended on macOS and Linux. Check with `viden --help`. Release archives exist for Windows and manual install.
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- The agent runs shell commands and writes files in the workspace, and does so without prompts when approvals are off.
- Delegated external agents are also launched from the cockpit.
This is third-party code. Review the repository files before installing.
What it does
Viden, a local coding agent written in Rust, opens a single cockpit in the terminal. It shows the transcript, approval state, a workspace snapshot, active tasks, diagnostics, provider health and recent evidence. The agent reads and edits files, searches the project, runs shell, Git, LSP and tests, and keeps tasks and memory. Every mutating action passes through permission modes. Models connect through one interface: DeepSeek, OpenAI, Anthropic, compatible gateways, Ollama and an offline fallback. External agents, Codex and Claude among them, appear in the cockpit as separate supervised lanes: shell and templates, tmux and an embedded PTY. Transcripts, tasks and memory are stored locally so work can resume.
Who it is for. Developers who coordinate work across several agents and want conversation, approvals and test evidence in one window.
Good fit when
- Codex and Claude work at the same time and need to be driven from one terminal.
- Every file edit, shell command or Git action must be approved before it touches the workspace.
- Work runs on a local Ollama model and an online provider is added only when needed.
- After a break a session must continue with stored tasks, memory and test evidence.
Not a fit when
- You need a plugin or skill for an agent already in use: Viden is an agent of its own, separate from yours.
- You need a mature project with a long release history: the version is 0.1.x.
- You work on a system outside the build list: macOS Apple Silicon and Intel, Linux x64, Windows x64.
Example request
Start Viden in this project, connect Ollama and propose a fix for the failing test. Show every file write for approval.Limitations
The project is early, with version 0.1.30 in the README. Builds exist for macOS on Apple Silicon and Intel, Linux x64 and Windows x64, and on macOS and Linux it installs through Homebrew. A clean install defaults to DeepSeek as the online provider, and others are added by hand in `/setup`. A mode without approvals can change files and run shell and Git, so keep a strict mode for the first runs. External ACP agent lanes are marked experimental.
How to disable. Run `brew uninstall viden` if you installed through Homebrew, or delete the `viden` binary from `PATH`. Delete local transcripts and memory by hand.
FAQ
How do I install Viden?
On macOS and Linux run `brew install wikieden/tap/viden`, then `viden --help`. For Windows there is a release archive and commands in the README.
Which models does Viden support?
DeepSeek, OpenAI, Anthropic, OpenAI-compatible gateways and Ollama, plus an offline fallback. The provider is chosen through `/setup` and `/models`.
How does Viden manage other agents?
It shows Codex, Claude and custom shell lanes as supervised lanes with state and evidence. Launch goes through tmux, an embedded PTY or a template.
What does the user approve?
File edits, shell commands, Git actions, workflows and delegated agent actions. The permission mode sets the level of control.
How do I open the command palette?
Press `Ctrl-P`. Provider and model setup is available through `/setup`, `/connect` and `/models`.
Discussion
No questions or reviews yet.
Related
CLI toolsA terminal coding agent built on Codex, tuned for low-cost open models like Kimi K3, DeepSeek and Qwen
A local control panel for coding agents: it runs Claude Code and Codex, keeps a task queue and merges verified work into main.
A terminal app that builds a team of Codex, Claude and Ollama agents and runs tasks in parallel under a lead agent
A terminal coding agent in one Rust binary that reads code, edits files and runs tests with any model of your choice.
A Rust CLI that builds a knowledge graph of code and decisions and feeds the agent the right slice through Claude Code hooks
Локальная среда для команды агентов: проекты как каталоги, промпты как файлы и журнал каждого шага, который можно проверить.
README in short
The author describes the project as a local operator cockpit for programming work, version 0.1.30. The README is available in English and Chinese.