Use Vector Studio like an engineering instrument.

Installation

Download Vector Studio for your platform and run it. No signup is required.

macOS

Download the DMG installer for your CPU, open it, drag Vector Studio to Applications, and launch it from Applications.

Windows

Download the EXE installer, run it, and follow the installation wizard. Vector Studio will be added to your Start menu.

Your first project

Open any folder. Vector Studio reads project metadata, source files, and git context to prepare an agent workspace.

Create a new project

1. Launch Vector Studio.
2. Choose a folder for the project.
3. Let Vector Studio initialize local project state.
4. Start with a short architecture or feature brief.

Open an existing project

Open the folder that already contains your code. Vector Studio can inspect package metadata, git history, and source files so the agent starts with concrete context.

Plan, execute, verify

1. Plan

Write a specification with requirements, constraints, and acceptance criteria. Ask the agent to validate the plan before changing files.

2. Execute

Let the agent edit files, run commands, inspect output, and iterate against the spec.

3. Verify

Review the diff and command output. If the implementation is correct, ship it. If it misses the spec, revise the instruction and continue from the evidence.

Spec-driven design

Instead of a vague request, give Vector Studio a target it can verify.

Build a login screen with:
- Email and password fields
- Remember me checkbox
- Clear errors for invalid credentials
- Loading state during submission
- Redirect to dashboard after success

Constraints:
- Use existing React form patterns
- Validate email before submit
- Add focused tests for failure and success paths

This works because the agent has a clear destination, and the review can be judged against the same written contract.

AI agents

Vector Studio orchestrates agent work from the desktop app. Conversations persist across sessions, and the agent can use local tools you approve.

• Read and edit project files.
• Run terminal commands for tests, builds, and scripts.
• Inspect git history and current diffs.
• Use MCP-connected services when configured.

Tools and MCP integration

Built-in tools cover files, terminal sessions, git, and project metadata. MCP support lets you add external systems where your workflow needs them.

• Database query tools for application state.
• API clients for product integrations.
• Custom analyzers, linters, and internal services.
• Enterprise systems exposed through your own MCP servers.

Code editor

The editor is built for reading and reviewing agent changes, with multi-file editing, syntax highlighting, search, and git diff views.

Terminal integration

Vector Studio detects common run targets such as tests, builds, and dev servers. Agents can execute commands and observe output in real time.

Chat interface

Chat is the control surface for directing the agent, reviewing tool use, and refining the spec. Use file references and concrete acceptance checks to ground the work.

AI providers

Vector Studio supports provider-backed workflows through your own configured API keys. Choose the provider that fits the task, then keep the same local execution loop.

• Claude: planning, coding, and long-running repository work.
• OpenAI: general coding, review, and tool-driven development.
• Gemini: an additional model route for project analysis and implementation.

Configuration

Vector Studio stores local project state in a .vector/ directory inside the project folder.

• .vector/state.json stores project metadata.
• .vector/chat/ stores chat history.
• .vector/logs/ stores event logs.
• .vector/.env stores provider configuration and secrets.