The AI agent that
knows your university.

Analyzes KPIs, simulates interventions, searches institutional documents, generates reports. One API call. You ask, it reasons.

Built on PydanticAI + Claude Sonnet · ANVUR-native · Source on GitHub →


curl
curl -X POST https://agent.didaflow.ai/v1/messages \
  -H "Authorization: Bearer df-unibo-a8f3..." \
  -H "Content-Type: application/json" \
  -d '{
    "message": "Why are students dropping Calculus I?",
    "program_id": 1
  }'

response

{
  "content": "Calculus I has the highest criticality index
    (33.5). Pass rate contribution: 76.9%. 11/30 students
    inactive. Recommended: Peer Tutoring [MODERATE]...",
  "tools": ["get_teaching_criticality", "recommend_actions"],
  "model": "anthropic:claude-sonnet-4-6",
  "cost_usd": 0.052
}

S01

KPI Analysis

36 ANVUR indicators at three levels. Cross-cohort comparison with significance.

S07

Impact Simulation

Pearl causal framework. Signed mediator deltas, confidence-weighted projections.

S06

Risk Prediction

LightGBM regularity model. Single, batch, and explainable predictions.

S04

Knowledge Search

RAG over regulations, syllabi, SUA, minutes. Voyage AI + Qdrant + FlashRank.

S08

Report Drafting

Riesame Ciclico, Schede Monitoraggio, Relazioni Paritetiche. Learns over time.

S02

Flow Graphs

Student flow as React Flow JSON. Stages, transitions, bottlenecks. Never individuals.



Add to your pipeline

Start free. Scale when the agent proves its value. No per-seat licensing. One API key per institution.

Sandbox

Try it now

Pre-loaded demo data
No API key needed
Rate-limited
See what the agent can do

Institution

Annual license

Unlimited programs
Unlimited requests
Dedicated instance available
SLA + priority support

Get started →

ANVUR IC01–IC28 EU data residency GDPR EU AI Act audit log Hetzner Helsinki