Guides
Conversational Care is open source. Here is where to find the code, how to get it running, and the ways we can work on it together.
✳
Three ways to get started
Run it yourself, have us alongside you, or build the next piece of research together — start wherever fits.
01 — Use it
Download and run it yourself
Clone the repository and follow the steps below: the full platform runs locally in a handful of commands. The technical manual covers architecture, configuration, and deployment; the user manual is written for care teams.
See the install steps02 — With us
Paid serviceHave us alongside you
Not every organisation wants to run a deployment alone. We can set the platform up in your setting, adapt the agents and workflows to how your team works, and take you through information governance and safety. We offer this as a paid service — it’s what keeps the open-source version maintained for everyone.
How it works03 — Research together
Partner with us on research
We’re looking for research partners to build on Conversational Care in grant applications — as a platform to deploy, evaluate, or extend. We bring the technology, the deployment experience, and the evidence so far, and we’ll support the application itself.
Start a conversation✳
Running it locally
The platform runs in Docker with PostgreSQL. These are the steps from the repository README — enough to have it up on your own machine in a few minutes.
Clone the repository
git clone https://github.com/palefo/conversational-care.git cd conversational-care/Django_CMSConfigure the environment
Set DJANGO_SECRET_KEY and whichever API keys you need.
cp .env.sample .envBuild and start the containers
Docker Desktop is the only prerequisite.
docker compose up -d --buildSet up the database
docker compose exec web python manage.py migrate docker compose exec web python manage.py createsuperuserOpen the platform
The site is at localhost:8000, and the admin at localhost:8000/admin.
Read the docs
Everything else — configuration, deployment, roles, agents, backups — lives in the repository.
Stuck, or found something? Open an issue on GitHub or get in touch.
✳
Having us alongside you
Not every organisation wants to run a deployment alone. This is how an engagement runs — the same process we used in the UK, Peru, and Singapore — and what you keep at the end of it.
Phase 01
User research & context
We map your specific needs together — your workflows, who you serve, and how you currently communicate with families.
Phase 02
Co-design
You decide what AI handles (like reminders or common questions) and what stays human-led (like complex decisions). We help you set it up your way.
Phase 03
Development & implementation
We integrate the system into your setup or build from scratch. You test with real families, we refine together, and you stay in control.
What we bring
- Deployment in your infrastructure or ours, with the Care Dashboard and agents configured for your setting
- Agents, workflows, and languages adapted to how your team actually works
- Information governance, safety, and escalation paths documented with you
- Training for care teams, and support while the first families come on board
What you keep
- Your deployment and your data — the platform runs where you run it
- Every change we make, published back to the open-source code
- No lock-in: if we step away, everything keeps working and any team can pick it up
This is a paid service, scoped to each deployment — and it is what keeps the open-source version maintained for everyone. Engagements usually start with a call to understand your setting.
Book a call