Get Started
DeepTutor ships four installation paths. They share one runtime-home layout: settings live in data/user/settings/ under the directory you launch from (or under DEEPTUTOR_HOME / deeptutor start --home if you set one explicitly). Knowledge bases, memory, Partner runtime state, multi-user accounts, and audit logs also live under the same data/ tree, so persist that tree and any separately configured content-workspace directories. The runtime home is the deployment storage root; it is distinct from the workspaces selected inside the app.
For the full app, the recommended flow is pick a runtime-home directory → install → deeptutor init → deeptutor start.
Pick a path
Section titled “Pick a path”| Option | When to use | Page |
|---|---|---|
| PyPI | Published packages; full app uses [full] in 1.7.0 | Install from PyPI |
| From Source | Current dev, local development, optional hot reload | Install from Source |
| Docker | Production deployments, hosted hosts, persistent volumes | Docker |
| CLI Only | Headless server, agent harness, no Web UI needed | CLI-Only Install |
1.7.0 installation:
deeptutor[full]installs the Web app; baredeeptutorinstalls core CLI/SDK. Stop DeepTutor before changing modules and keep feature wheels at the core version. For development, use the source instructions or build local wheels. Back up your runtime home before upgrading older installations.
Minimum requirements
Section titled “Minimum requirements”| Component | Version |
|---|---|
| Python | 3.11–3.14 |
| Node.js | 20+ (PyPI) / 22 LTS (from source, matches CI + Docker) |
| RAM | Depends on the selected models and workload |
| Disk | Depends on install extras and workspace data |
You’ll also need an LLM provider. Hosted services usually require an API key; local providers such as Ollama, LM Studio, llama.cpp, and vLLM can run without one. An embedding provider is optional unless you plan to use Knowledge Bases.
What gets bound
Section titled “What gets bound”After deeptutor start, two services come up:
| Service | Default port |
|---|---|
| FastAPI backend | 8001 |
| Next.js frontend | 3782 |
Both are overridable via data/user/settings/system.json or during deeptutor init.
After install
Section titled “After install”Once you can reach http://localhost:3782/:
- Explore DeepTutor — start with workspaces, then Home, Learning, Resources, Partners, and Personalization
- DeepTutor CLI — drive core runtime and workspace operations from the terminal
- Partners: channels — connect autonomous tutors to WeChat, WeCom, QQ, Telegram, Slack, Feishu, and more
- Multi-User Deployment — host DeepTutor for a team
If something breaks, jump to Troubleshooting.
