Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/AsyncFuncAI/deepwiki-open/llms.txt

Use this file to discover all available pages before exploring further.

Manual setup runs DeepWiki Open as two separate processes on your machine: a FastAPI backend (Python 3.11+) that handles AI inference and repository analysis, and a Next.js frontend that serves the browser UI. This approach gives you full control over each layer and is the recommended path for local development.
You need at least one API key to generate wikis. Provide GOOGLE_API_KEY for Google Gemini models or OPENAI_API_KEY for OpenAI models. If you want a fully offline setup with no API keys at all, use Ollama instead.

​
Prerequisites

RequirementVersionNotes
Python3.11+The .python-version file in the repo pins 3.12
Poetry2.0.1Installed via pip in the setup steps below
Node.js20+Required for the Next.js frontend
npm or Yarn—Comes with Node.js; Yarn is optional

​
Setup steps

1

Clone the repository

git clone https://github.com/AsyncFuncAI/deepwiki-open.git
cd deepwiki-open
2

Create a .env file

Create a .env file in the project root. Only one cloud provider key is strictly required to generate wikis, but you can add as many as you want.
# --- Required: at least one of these ---
GOOGLE_API_KEY=your_google_api_key
OPENAI_API_KEY=your_openai_api_key

# --- Optional: additional providers ---
OPENROUTER_API_KEY=your_openrouter_api_key

# Azure OpenAI
AZURE_OPENAI_API_KEY=your_azure_openai_api_key
AZURE_OPENAI_ENDPOINT=your_azure_openai_endpoint
AZURE_OPENAI_VERSION=your_azure_openai_version

# Ollama (only if Ollama is on a remote host; defaults to http://localhost:11434)
OLLAMA_HOST=your_ollama_host

# --- Optional: embeddings provider ---
# Options: openai (default), google, ollama
DEEPWIKI_EMBEDDER_TYPE=google

# --- Optional: server config ---
PORT=8001
SERVER_BASE_URL=http://localhost:8001
3

Install Python dependencies

python -m pip install poetry==2.0.1 && poetry install -C api
This installs Poetry and then uses it to install all backend dependencies declared in api/pyproject.toml into an isolated virtual environment inside api/.
4

Start the backend API server

python -m api.main
The FastAPI server starts on port 8001 by default. You will see log output confirming it is ready to accept requests. Leave this terminal open.
5

Install frontend dependencies

Open a second terminal window in the project root, then install the JavaScript packages:
npm install
6

Start the Next.js frontend

npm run dev
The frontend starts on port 3000 in development mode with hot-reloading enabled.
7

Open DeepWiki in your browser

Navigate to http://localhost:3000. Enter any GitHub, GitLab, or Bitbucket repository URL and click Generate Wiki.For private repositories, click + Add access tokens and provide your personal access token before generating.

​
Default ports

ServicePortEnvironment variable
FastAPI backend8001PORT
Next.js frontend3000—
If you change PORT, also update SERVER_BASE_URL so the frontend knows where to find the API: 
PORT=9000
SERVER_BASE_URL=http://localhost:9000

​
Choosing an embedder

The embedder controls how repository code is indexed for retrieval. Set DEEPWIKI_EMBEDDER_TYPE in your .env file:
ValueEmbedding model usedAPI key required
openai (default)text-embedding-3-smallOPENAI_API_KEY
googlegemini-embedding-001GOOGLE_API_KEY
ollamanomic-embed-text (local)None
When you switch embedders, previously generated embeddings from other providers are incompatible. You will need to re-index any repositories you had already processed.

Build docs developers (and LLMs) love

Get started for freeTalk to us