AI Safety
Put policy and emergency controls around autonomous actions.
Open AI infrastructure from Sarajevo
Building safer AI systems people can trust.
We build practical infrastructure that helps organizations control autonomous agents, use trusted knowledge, measure AI behavior, govern tools and apply AI to real-world work.
6published systems
5+live web apps
DOIpermanent research record
33%humanitarian commitment
AITRUSTGOVERN · MEASURE · CONTROL
🛡️Protect
⌕Ground
◈Measure
🔐Control
About Us
Welcome to We Do Care Global
We Do Care Global is an autonomous AI governance initiative founded by Emir Perla, AI Workflow & Automation Engineer based in Sarajevo, Bosnia and Herzegovina. We build open-source, production-ready AI systems that combine policy engines, retrieval proofs, measurement platforms, governed tool execution, and continuously-archived research infrastructure.
6Published Core Systems
5+Live Web Applications
10Permanent Zenodo DOIs
33%of Revenue to Humanitarian Causes
All systems are fully open-source under Apache-2.0, with dedicated live webapps, tamper-evident GitHub repositories, and permanent CERN Zenodo DOIs. Our mission is to ensure AI serves humanity through transparent, accountable, and beneficial technology.
What we actually do
Five clear jobs. One ecosystem.
Instead of one giant AI dashboard, WDCG is a set of focused systems that can work together or stand alone.
Trusted Knowledge
Give AI better enterprise retrieval, ranking and evidence.
Evaluation
Measure reliability, latency, tool behavior and outcomes.
Tool Governance
Control permissions, secrets and AI-to-tool execution.
Applied Intelligence
Turn the stack into useful multi-agent and regulatory workflows.
The portfolio
Six systems. Six jobs.
Open the live product, source repository or permanent research record. Status labels are intentionally simple.
🛡️LIVE
AgentGuard
Protects AI agents before they can take risky or unauthorized actions.
🏛️PUBLISHED
FoundationCentar
Coordinates multi-agent AI work with bounded workers, human approval and evidence checks. Try the interactive demo.
⌕LIVE
Enterprise Hybrid-RAG
Helps AI find better enterprise information by combining retrieval methods, reranking and evaluation.
◈LIVE
Agent Eval
Measures whether AI agents are reliable: behavior, latency, tool use and benchmark performance.
🔐LIVE
AgentVault
Controls which AI tools can be used, which permissions they receive and how activity is tracked.
💊LIVE
Pharma Intelligence OS
Focused pharmacovigilance and regulatory intelligence workflows with human verification.
How value is created
From AI idea to trusted system.
Follow the stack from protection to application. Each layer can be used independently.
01Protect
→
AgentGuard
02Ground
→
Hybrid-RAG
03Measure
→
Agent Eval
04Control
→
AgentVault
05Apply
FoundationCentar · Pharma OS
Evidence Center
Unify GitHub, ORCID, Zenodo, releases, architecture notes and verification results so researchers, partners and buyers can verify the work quickly.
See evidence →Pilot → Production
Give organizations one simple route: choose a problem, run a bounded pilot, measure the result, then deploy the governed system.
Start a pilot conversation →
Evidence & research
Open work. Permanent record.
WDCG keeps the public proof chain visible: source code, authorship, licensing and versioned DOI records.
33%Humanitarian commitment
Mission
Technology should create human value.
We Do Care Global carries a founder-level commitment to dedicate 33% of annual corporate revenue to direct financial and material humanitarian assistance.