Talimio

An all-in-one, AI-powered learning platform that adapts to you.

Why Talimio?

• Turn any topic into a full course in seconds. Upload your own documents to ground the AI with RAG, or let the model build from scratch.
• Create adaptive courses where the AI builds a dependency graph (DAG) and only unlocks new lessons once you’ve mastered the prerequisites.
• We don’t do boring flashcards. A LECTOR-based scheduler tracks concepts you struggle with and resurfaces them in future lessons exactly when you need them.
• Lessons are fully interactive. Like Claude Artifacts or the demos on Brilliant, you get hands-on widgets you can click and play with right inside the lesson.
• Executable code blocks are built in: run, tweak, and break code directly inside the lesson.
• Answer quick questions and self-assessment quizzes inside lessons; your responses continuously shape and adapt the course to you.
• Upload a PDF book or drop in a YouTube link and chat with it instantly.
• Run the whole stack 100% offline with Ollama, or plug in your own cloud API keys (supports many providers via LiteLLM).
• Persistent memory of your preferences and learning style, so it gets better at adapting to you the more you use it.
• Extensible via the Model Context Protocol (MCP), so you can connect your own tools and APIs to interpret data or take actions.

Quick Start (Docker Compose)

Prerequisites: Docker and Docker Compose installed.

Environment: All required .env values are defined directly in docker-compose.yml. You do not need backend/.env. To customize, edit the backend service environment block and optionally uncomment provider API keys.

1. Clone the repo

git clone https://github.com/SamDc73/Talimio.git
cd Talimio

2. Start the stack

docker compose up -d

3. Open the apps

• Frontend: http://localhost:5173
• API (FastAPI): http://localhost:8080

4. Optional: Pull Ollama models (first run)

• Skip if you use cloud LLMs and set provider keys in docker-compose.yml.

docker exec -it ollama ollama pull gpt-oss:20b
docker exec -it ollama ollama pull nomic-embed-text

Note: To disable the local LLM, comment out the ollama service in docker-compose.yml and set PRIMARY_LLM_MODEL to a cloud model. Provide the relevant API key(s) in the same environment block.

5. Stop/Update the stack

docker compose down
# update later
docker compose pull && docker compose up -d

Local Development (without Docker)

1. Clone the repo

git clone https://github.com/SamDc73/Talimio.git
cd Talimio

2. Backend (Python 3.12+, uv):

cd backend
uv sync
cp .env.example .env
uv run uvicorn src.main:app --reload --port 8080

3. Frontend (Node + pnpm):

• Now in a diffrent tab/window:

cd web
cp .env.example .env
pnpm install
pnpm dev

4. Now you can open the apps:

• Frontend: http://localhost:5173
• API (FastAPI): http://localhost:8080

Contributing

Any type of contribution is greatly apprietiated!

First timer flow:

1. Pick one tiny improvement

• Fix a typo, clarify a log/error message, tidy a function name, or improve copy.
• Good places: README.md, backend/src/** (messages/docs), web/src/** (copy/UI nits).

2. Run it locally

• Use Quick Start above, and test it out.

3. Run checks before PR

• Backend: cd backend && ruff check src --fix
• Backend types: cd backend && uvx ty check src
• Frontend: cd frontend && pnpm run lint

4. Open a draft PR

• Two bullets are enough: why it helps, what changed.

Support

Questions, help, or feedback? Join our Discord