๐Ÿ“ Text Modality

Meal Planning & Nutrition

Budget-constrained meal plans tagged by household demographics, dietary restrictions, region, and season. Each entry cross-referenced with USDA nutritional data and validated by a Registered Dietitian Nutritionist.

๐Ÿฅ—
Budget Meal Plan #4,217
Meal Planning & Nutrition ยท hmd_meal_004217
โœ“ RDN Verified
household_size4 (2 adults, 2 children ages 6 & 9)
weekly_budget$85.00 USD
dietary_tagsdairy-free nut-safe
regionSoutheast US
seasonsummer
income_band$45,000โ€“$65,000/yr
avg_daily_calories1,820 kcal / adult
formatJSON Q&A pairs Knowledge Graph
expert_validatedโœ“ RDN โ€” Sarah Okonkwo, MS, RDN
๐ŸŽ™๏ธ Voice/Audio Modality

Voice Commands (Smart Kitchen & Home)

Multi-accent wake-word samples, appliance control phrases, ambient kitchen noise overlays, and domain-specific command utterances. Each entry annotated with room classification, ambient noise label, and command intent taxonomy. Premium category for smart speaker and home AI teams.

๐ŸŽ™๏ธ
Voice Command Sample #892
Voice Commands ยท hmd_voice_000892
Audio
command_text"Set a timer for the roast for 45 minutes"
intentkitchen_timer_set
accent_regionAppalachian English (Southeast US)
speaker_genderfemale
ambient_noisestove_active_low
room_classkitchen
snr_db18.4 dB
duration_ms2,340 ms
audio_formatWAV 16kHz transcript + intent labels
copyright_cleanโœ“ consent-documented collection
๐Ÿ“ Text Modality

Household Economics

Budget frameworks, grocery benchmarks, energy cost data, and financial decision trees โ€” tagged by income band, region, and household composition. Validated by Certified Financial Counselors.

๐Ÿ’ฐ
Household Budget Framework #1,089
Household Economics ยท hmd_econ_001089
โœ“ CFC Verified
household_size3 (2 adults, 1 child age 4)
annual_income$58,000 USD
income_band$45,000โ€“$65,000/yr
regionMidwest US
housing_typeRenter โ€” 2BR apartment
monthly_housing_cost$1,140 (29.7% of gross)
monthly_grocery_benchmark$520 (USDA moderate-cost plan)
savings_rate8.4% of net income
expert_validatedโœ“ AFCยฎ โ€” Marcus Webb, CFC

Schema Reference

Core fields shared across all dataset categories. Every entry conforms to this base schema, with category-specific and modality-specific extension fields layered on top. Built for direct ingestion into RAG pipelines, fine-tuning workflows, and evaluation frameworks.

entry_id
string
Unique identifier. Format: hmd_{category_code}_{id}. Stable across dataset versions.
e.g. "hmd_meal_004217", "hmd_voice_000892"
category
string
Dataset category. One of 11 defined values covering all domestic knowledge domains.
e.g. "meal_planning_nutrition", "voice_commands_smart_home"
modality
string
Data modality: text, audio, vision, or multimodal. Enables modality-specific filtering at query time.
e.g. "text", "audio", "vision", "multimodal"
household_profile
object
Household demographic data: size, composition, income band, region, and cultural context. Enables precise demographic filtering.
e.g. { "size": 4, "income_band": "45000_65000", "region": "southeast_us" }
expert_validation
object
Validation metadata: boolean flag, validator credential type (RDN, AFCยฎ, CFC, GC), validation date, and revision history. Every entry has validated: true before inclusion.
e.g. { "validated": true, "validator_credentials": "MS, RDN" }
provenance
object
Origin and licensing metadata. origin is always "original_mhai". copyright_clean is always true. Includes license version string for legal audit trails.
e.g. { "origin": "original_mhai", "copyright_clean": true }
optimization_flags
array
Machine-readable flags identifying optimization opportunities within the entry context. Useful for building recommendation systems on top of the dataset.
e.g. ["energy_reduction_opportunity", "childcare_high_pct"]
qa_pairs
array
Array of 2โ€“6 question/answer objects derived from the structured entry. Pre-formatted for fine-tuning and RAG retrieval. Each pair includes context metadata.
e.g. [{ "q": "How should...", "a": "For a family...", "tags": [...] }]
decision_tree_ref
string
Reference ID to an associated decision tree document (Professional+ plans). Links structured data to branching reasoning logic for RAG and agentic workflows.
e.g. "hmd_dt_budget_renter_midwest_v2"

Ready to go deeper?

Download 500 free entries across two categories, or get in touch to discuss your specific data and modality needs.