01 / UnderstandIntent
Express the network you need.
Start from engineering intent instead of manually assembling disconnected device commands.
Live workspace available
/
OpenAI Devpost entry ↗
AI-native Network Engineering Platform
Network Engineering.
Without the Guesswork.
Llebre turns network intent into reviewable, vendor-ready output—and validates the architecture before it reaches production.
Launches at app.llebre.com · no installation required
Input
Network intent
Output
Vendor-ready CLI
Evidence
Architecture validation
Product interface
Network IDE / Vanguard
app.llebre.com
> design campus core
✓ Intent parsed
✓ Policy checks passed
→ Ready for review
State
Reviewable
A different operating surface
Not a chatbot.
A network engineering workflow.
Llebre keeps intent, target profile, generated artifacts and validation in the same workspace. The result is visible, reproducible and designed for review—not hidden inside a conversation.
02 / ProduceArtifacts
Generate output worth reviewing.
Target a vendor profile and turn the plan into explicit configuration artifacts, not prose.
03 / ProveArchitecture
Validate before production.
Expose inconsistent intent, policy problems and architectural risk before deployment.
Infrastructure deserves engineering. Not autocomplete.
From intent to production
One continuous engineering loop.
Every stage leaves something inspectable behind.
01
Design
Define topology, services and constraints.
02
Validate
Check intent and architectural consistency.
03
Generate
Create vendor-targeted artifacts.
04
Review
Inspect output, assumptions and findings.
05
Deploy
Move forward with evidence.
06
Monitor
Track drift, health and compliance.
OpenAI inside the engineering loop
AI handles ambiguity.
Engineering handles truth.
OpenAI-powered reasoning interprets the engineer’s intent and helps structure a plan. Llebre then turns that plan into explicit artifacts and runs validation against the architecture.
The model is part of the workflow—not the final authority. Generated output remains visible, inspectable and subject to deterministic checks.
Model layer
Interpret intent
Resolve natural-language requirements into a structured engineering plan.
Artifact layer
Produce output
Generate configurations and operational artifacts for the selected target.
Evidence layer
Validate claims
Run architectural checks and surface findings before deployment.
Built for the OpenAI Devpost hackathon
Competition page ↗
Interactive proof
Understand Llebre in under a minute.
Run the guided preview, then open the real workspace.
01Intent parsed
02Artifact generated
03Risk exposed
Guided evidence preview
This preview does not replace the live app
Terminal outputTarget / EOS
$ llebre plan campus-core --target arista-eos
Ready. Run the architecture preview.
The product
A network IDE. Not another SaaS dashboard.
One focused surface for the command deck, target profile, generated CLI and architecture review.
Actual Llebre interfaceLocal-first / deterministic
Why Llebre
Fast because the path is clear.
Llebre means hare in Valencian. The name carries speed, awareness and the ability to change direction without losing control—exactly what modern network engineering demands.
The brand is Mediterranean. The product is built for network engineers everywhere.
Live project · testable now
Stop configuring.
Start engineering.
App / app.llebre.com
Updates / x.com/llebrecom ↗