Overview
Research-grade deep dives into what 20 years of GOAP literature has taught us about feel, performance, and stability — and how Intent Forge applies those lessons.
[ Deep Dives ]
The why behind every design choice
For readers who want to understand the why — what twenty years of GOAP literature, AAA postmortems, and academic work has taught us about making planner-driven AI feel right. Intent Forge's design choices aren't arbitrary; these pages trace the heritage and explain the tradeoffs we accepted (and the ones we rejected).
Pick your dive
Anti-Flap Toolkit
The flagship. Five-family stability layer, the 20-year flapping literature behind it, and validator evidence for each layer.
Planner algorithms
Why A* — and where Dijkstra, Weighted A*, Anytime A*, D* Lite, and HTN sit relative to our use case.
HTN vs GOAP
Both paradigms ship AAA games. When each one wins, when hybrids work, and why Intent Forge is GOAP.
Networking & replication
Server-authoritative GOAP with replicated outputs. The shipped v1.0-alpha design, what crosses the wire, and the v2.0 client-prediction candidate.
Ecosystem & Positioning
Honest 2026 UE5 AI landscape, who Intent Forge is for, and where it fits in the LLM-shaped future. Named competitors, AAA shipping record.
References
The full bibliography — Orkin, Jacopin, Conway, Girard, Boeda, Marko, Smrček, Decima, Aversa, Game AI Pro, GDC, AAAI.
How these pages are written
These are not marketing pages. They're written in a research-paper tone: claims are cited, alternatives are taken seriously, and where we made a tradeoff we say so. If you find a citation you disagree with or a paper we should have read, open an issue — the bibliography is meant to grow.