DeepWiki Open
AsyncFuncAI/deepwiki-openReviewed by AI
Checked by RepoPicks Radar on
What is DeepWiki Open?
DeepWiki-Open reads a repository you name and uses an AI model to write documentation about it: a structured wiki, Mermaid diagrams of how the parts fit together and a code map for guided tours of the code. You run it on your own machine with Docker Compose (a web page on port 3000 plus an API on 8001) or on Kubernetes with the included Helm chart, and you supply your own model keys — OpenAI and Google by default, with OpenRouter or a local Ollama model as alternatives. The author calls it "my own implementation attempt of DeepWiki".
How easy is it to install DeepWiki Open?
D · terminal, several tools or a specific graphics card · Windows
- Windows D
- macOS D
- Linux D
Same route and same grade on the three systems: install Docker, write a .env file with your OpenAI and Google keys, build and start the container with docker-compose.yml and open port 3000. The compose file reserves up to 6 GB of memory, and the only step-by-step in the README is the Kubernetes one with kubectl and Helm.
Coming soon: install in 1 click with the RepoPicks app.
DeepWiki Open on GitHub
- Stars
- 18,397on Nov 10, 2026
- Downloads
- —all releases
- Growth
- +28stars in 6 days
Stars over time
Facts
- Runs on
- Windows, macOS, Linux
- You need to install
- Docker
- To build from source
- Node.js, Docker
- Disk space
- 100 MB–500 MB · estimate
- Accounts or keys
- OpenAI, Google (Gemini), OpenRouter or a local Ollama model as an alternative to the two above
- Cost
- Free, API costs extra · The code is MIT and free. By default you pay OpenAI and Google per repository you index; the repository also ships a Dockerfile and instructions for using a local Ollama model instead.
- License
- MIT (Permissive)
- Last change on GitHub
- Nov 10, 2026
DeepWiki Open risk signals (5)
- Attentiondefault passwords in the litellm setup
docker-compose-litellm.yml publishes ports 4000 and 5432 to the host and falls back to fixed secrets if you do not set them: "LITELLM_MASTER_KEY=${LITELLM_MASTER_KEY:-sk-1234}" and "POSTGRES_PASSWORD: ${LITELLM_DB_PASSWORD:-litellm_password}".
AI review
- Infoyour code is sent to the model provider
The container warns "OPENAI_API_KEY and/or GOOGLE_API_KEY environment variables are not set. These are required for DeepWiki to function properly" — to write the wiki the tool sends the repository's code and text to those companies, which matters for private repositories unless you switch to the local Ollama setup.
AI review
- Infonewer version is offered on an outside site
README section "Deepwiki-Open 2.0 (Grok Wiki is now live)" says only "Download at https://grok-wiki.com", which is also the repository homepage; what that site delivers is not in this repository.
AI review
- Inforeadme dropped the local install steps
The README has no quick start for running it locally — the only commands shown are kubectl and helm — even though docker-compose.yml, Dockerfile, Dockerfile-ollama-local and run.sh are in the repository.
AI review
- Infoneeds a lot of memory
docker-compose.yml sets "mem_limit: 6g" and "mem_reservation: 2g" for the container, and the Next.js build runs with NODE_OPTIONS="--max-old-space-size=4096".
AI review
Evidence is shown in its original language.
What Radar could not check about DeepWiki Open
- How much the OpenAI and Google calls cost for one repository, which depends on its size
- How good and how correct the generated wiki and diagrams are
- Whether the local Ollama setup gives usable results without the paid keys
- Real disk use in ~/.adalflow after indexing a few repositories
- Whether the program at grok-wiki.com is built from this same code