Installation

Vectora ships as a native desktop app (Electron + a compiled Python backend) for Windows, macOS, and Linux, with built-in auto-update. There is no “single Docker image” of the product — Docker, when used, only spins up optional infrastructure (Postgres/Redis/Qdrant for complete mode), never Vectora itself.

Option 1 — Native installer (recommended)

Download the installer for your OS:

OS	Format	Signing
Windows	`.msi` or `.exe` (NSIS)	EV certificate (Azure Trusted Signing)
macOS	`.dmg` (Apple Silicon only)	Apple Developer ID + notarized
Linux	`.AppImage`, `.deb`, or `.rpm`	unsigned

Install normally (double-click / dpkg -i / rpm -i). The app opens with the backend already embedded — no separate Python, Node, or any other dependency to install.

Future updates arrive automatically via auto-update (served by updates.vectora.company).

Option 2 — From source (dev)

To contribute or run in development mode:

Requirements: Python 3.13, uv, Node.js 24+, pnpm.

git clone https://github.com/vectora-company/vectora.git
cd vectora/vectora

uv sync                          # Python dependencies
pnpm --dir frontend install       # frontend dependencies

cp .env.example .env
# edit .env: GOOGLE_API_KEY (or another provider), COHERE_API_KEY, TAVILY_API_KEY

Two terminal windows:

# Terminal 1 — full backend + SPA (port 8080)
uv run vectora start --port 8080

# Terminal 2 — frontend dev server (Vite, port 3000, proxies to the API)
pnpm --dir frontend dev

Open http://localhost:3000. The first user to sign up becomes the root administrator.

Required API keys

Key	Required?	For
An LLM provider (Gemini, OpenAI, Anthropic, Cohere, or local Ollama)	Yes	Chat, code generation, response synthesis
`COHERE_API_KEY` (or VoyageAI)	Yes, for RAG	Embeddings + reranking
`TAVILY_API_KEY`	Optional	Web search

The model selector in chat only shows providers with a configured key — no key, no provider in the list.

License

The app works without a license in Free mode (100% local). To unlock Pro features (multi-user web chat, complete storage, webhook-triggered automations), you need a VECTORA_TOKEN, obtained from the dashboard after subscribing to a paid plan.

Next step

→ Quick start

Introduction Quick Start