HealthChain
Danger signs first. Then where to go.
WHO danger-sign decision support and nearby facilities for community health workers.
Agent for Good Hub
Ask one question, in your own language. The Hub sends it to the right specialist — for health, education, water, energy, food, climate, credit, work, crisis response, endangered languages, and fairness in hiring and AI — and brings back a short answer first, then the detail.
12 specialist agents, all live, and the coordinator that routes between them. Each entry says what the agent does.
Routes every question
One question goes to the Hub coordinator. The coordinator picks the agent or composes several — and streams the answer back as it lands.
Ask the coordinatorDanger signs first. Then where to go.
WHO danger-sign decision support and nearby facilities for community health workers.
Draft a lesson that fits the learner.
Lesson plans and easier-to-read text for learners left out by mainstream tools.
Find the biased line. Get the neutral rewrite.
Checks job ads, hiring numbers and policy text for gendered bias.
Check the water point. Treat the water safely.
Looks up nearby water points and gives WHO household water-treatment steps.
Plan the day. Hold the evening peak.
Demand, dispatch and fault-triage estimates for off-grid solar and wind microgrids.
See where your qualifications could stand across the border.
Maps a refugee’s qualifications to host-country frameworks and livelihood pathways.
Check the surplus is safe. Plan who could use it.
Food-safety and matching checks for getting surplus food to kitchens and food banks.
Sketch the climate plan. No live data yet.
How heat, flood, tree-canopy and carbon questions are answered from satellite data.
Sort the reports. Draft the briefing.
Structures crisis reports and drafts 5W situation reports for people to decide on.
See how a fair credit decision should be explained.
Explains credit decisions in plain language, factor by factor, with appeal rights.
Plan the recordings. Keep the language yours.
Guidance on documenting an endangered language, with consent and community ownership.
Test the algorithm. Show the numbers.
Fairness statistics and audit outlines for AI systems, from the figures you provide.
Menus and the chat screen are in five languages; other pages are in English. Answers follow the language you write in — we have tested Swahili, Arabic and Hindi, and quality varies. The Hub runs in the browser and needs a connection. WhatsApp, SMS, USSD and voice are planned, not yet available.
One question goes to the Hub coordinator. The coordinator picks the agent or composes several — and streams the answer back as it lands. Twelve specialists cover health, education, gender equality, water, energy, work, food, climate, crisis response, credit, language and fairness in AI.
Danger-sign triage follows published WHO rules and is decision support, not diagnosis. Answers are asked to state their confidence, to say so when it is low, and to say when a figure is illustrative. The Hub cannot connect you to a person, and nothing is sent or booked for you: in an emergency, contact local emergency services.
We count only outcomes that a named partner, evaluator or community witness has attested. None has been attested yet, so we show no impact numbers. We are looking for partners to evaluate the Hub.
Read the methodCommunities know their context best. Sketch a WhatsApp, SMS, USSD or voice flow and export it as JSON. This is a drafting tool: nothing is published, connected to an agent or sent to anyone yet.