Base URL on the main site: https://thataiguy.org/api/models
Dedicated subdomain when DNS is finished: https://api.thataiguy.org/models
| Tier | Rate limit | Notes |
|---|---|---|
| Free | 10 searches/day | Browser demo and light testing |
| Basic | 1000 searches/day | Full recommendation payloads |
| Pro | Unlimited | Batch usage and priority support |
API key enforcement can be layered in through Nginx or Go middleware without changing the request shape.
Recommend models from a plain-English description.
{
"description": "what you need the model for",
"task_type": "text-gen|image-gen|code|embed|classify|vision|audio|multimodal|other",
"max_size_gb": 10,
"hardware": "cpu|gpu-8gb|gpu-24gb|apple-silicon",
"license": "open",
"top_k": 5
}
curl -X POST https://thataiguy.org/api/models/recommend \
-H "Content-Type: application/json" \
-d '{
"description": "generate realistic images from text",
"hardware": "apple-silicon",
"top_k": 5
}'
{
"results": [
{
"model": {
"name": "PixArt-alpha/pixart_sigma_sdxlvae_T5_diffusers",
"family": "pixart",
"task_types": ["text-to-image"],
"source": "huggingface"
},
"score": 0.82,
"reasons": [
"Matches your description semantically",
"Aligned with task: text-to-image"
]
}
],
"query_time_ms": 23,
"models_searched": 4002,
"models_after_filter": 258
}
Shortcut GET form for quick searches.
curl "https://thataiguy.org/api/models/recommend?q=small+language+model+for+code+completion+on+cpu"
Returns normalized model families with counts.
curl https://thataiguy.org/api/models/families
Returns task buckets with counts.
curl https://thataiguy.org/api/models/tasks
Returns models in a family.
curl https://thataiguy.org/api/models/family/qwen
Returns the full model entry and hardware estimates.
curl "https://thataiguy.org/api/models/detail/PixArt-alpha%2Fpixart_sigma_sdxlvae_T5_diffusers"