Back to Ecosystem Pulse
ToolExperimental
cidadao.ai-backend
by anderson-ntlabs
Multi-agent backend that turns transparency portal data into intelligent investigations
Python
Updated Sep 3, 2026
Share:
Summary
Transforms raw public transparency portal data into structured, investigable findings using a multi-agent AI backend. Orchestrates specialized agents (ingest, NLP, fact-checking, summarization) to turn open-data streams into human-readable investigations and alerts. Built as a Python/FastAPI service tailored for civic-tech workflows and large public datasets.
Why It Matters
As civic AI workflows scale, knowing which agent produced which claim and why matters for accountability. This project operationalizes agent delegation for public-data investigations, making provenance and structured outputs easier to audit. Provenance and structured outputs help surface agent failure modes and create trust signals or continuous evaluation later.
Best For
Civic-tech teams and researchers building automated investigations from government open-data who need an extensible agent pipeline.
Applications
- Automating ingestion and normalization of government transparency portal datasets
- Running specialist agents for fact-checking, summarization, and evidence extraction
- Producing human-readable investigation reports and alerts from open-data streams
Works With
fastapipython
Topics
aibrazilcivic-techfastapigovernmentmachine-learningmulti-agent-systemnlpopen-datapython+2 more
Similar Tools
autogencrewai
Keywords
multi-agent orchestrationcivic-techagent delegationopen-data