Skip to content
Unified Product GraphBeta
Get Started
Unified Product Graph

The open specification for product knowledge.
312 types. 36 domains.
MIT Licensed

Understand
The SpecBlog
Explore
Graph ExplorerEntitiesEdgesLifecyclesFrameworksPlaybooksApproaches
Adopt
MCP ServerSDKCLISkillsIntegrationsGitHub ↗
© 2026 Unified Product Graph · MIT License · Privacy · TermsBuilt by The Product Creator

Entities

312 entity types
⌕
🤖AI Model🗨️Prompt Version🎖️Eval Benchmark🎟️Eval Run💸AI Cost Tracker👻Hallucination Report🚦AI Guardrail🔄Model Comparison👾AI Experiment🪆AI Dataset🪡AI Trace✒️Prompt Template
‹
Reference/AI & Machine Learning25/36
›

🤖 Domain

AI & Machine Learning

The intelligence layer. AI models, prompts, evaluations, guardrails, and cost tracking.

12Entity Types
Part of region07Engineering & Platform↗
🤖
AI Model
An AI or ML model used within the product
12 props10 edges
→🗨️
Prompt Version
A version of a prompt template
8 props1 edges
→🎖️
Eval Benchmark
A benchmark for evaluating AI quality
4 props3 edges
→🎟️
Eval Run
An evaluation run against a benchmark
10 props1 edges
→💸
AI Cost Tracker
A tracker for AI inference costs
9 props1 edges
→👻
Hallucination Report
A report of an AI hallucination
5 props1 edges
→🚦
AI Guardrail
A guardrail for AI safety
3 props1 edges
→🔄
Model Comparison
A comparison between AI models
3 props2 edges
→👾
AI Experiment
An AI-focused experiment
11 props1 edges
→🪆
AI Dataset
A dataset for AI training or evaluation
9 props1 edges
→🪡
AI Trace
An AI inference trace
11 props3 edges
→✒️
Prompt Template
A reusable prompt template
3 props2 edges
→
‹
Reference/AI & Machine Learning25/36
›

🤖 Domain

AI & Machine Learning

The intelligence layer. AI models, prompts, evaluations, guardrails, and cost tracking.

12Entity Types
Part of region07Engineering & Platform↗
🤖
AI Model
An AI or ML model used within the product
12 props10 edges
→🗨️
Prompt Version
A version of a prompt template
8 props1 edges
→🎖️
Eval Benchmark
A benchmark for evaluating AI quality
4 props3 edges
→🎟️
Eval Run
An evaluation run against a benchmark
10 props1 edges
→💸
AI Cost Tracker
A tracker for AI inference costs
9 props1 edges
→👻
Hallucination Report
A report of an AI hallucination
5 props1 edges
→🚦
AI Guardrail
A guardrail for AI safety
3 props1 edges
→🔄
Model Comparison
A comparison between AI models
3 props2 edges
→👾
AI Experiment
An AI-focused experiment
11 props1 edges
→🪆
AI Dataset
A dataset for AI training or evaluation
9 props1 edges
→🪡
AI Trace
An AI inference trace
11 props3 edges
→✒️
Prompt Template
A reusable prompt template
3 props2 edges
→