API Reference

REST API for querying models, providers, and capabilities
Base URLhttps://api.ai-model.dev/v1
FormatJSON (append ?pretty for formatted output)

Overview

GET/v1/statsRegistry statistics: provider, model, family, creator counts.
GET/v1/searchUnified search across providers and models.
qSearch query (min 2 chars, required)
limitMax model results (default 20, max 50)

Providers

GET/v1/providersAll providers with model counts.
GET/v1/providers/:idSingle provider with full details and models.

Models

GET/v1/modelsList models. Filterable, sortable, paginated.
providerFilter by provider id
familyFilter by model family
creatorFilter by original creator
statusactive | deprecated | preview
capabilityreasoning | vision | tool_call | streaming | structured_output | json_mode | fine_tuning | batch
qSearch across id, name, description
sortname | context_window | price_input | price_output
orderasc (default) | desc
limitMax 500, default 100
offsetDefault 0
GET/v1/models/:provider/:idSingle model details. Supports / in IDs.
GET/v1/models/compareCompare up to 10 models side by side.
idsComma-separated provider/model IDs (2–10, required)

Taxonomy

GET/v1/familiesModel families with counts and metadata.
GET/v1/capabilitiesAll capabilities with model and provider counts.

Response format

Success{ "data": ..., "meta": { "total", "limit", "offset" } }
Error{ "error": { "message", "status" } }

Conventions

Field presentKnown value
Field omittedUnknown
Field is nullNot applicable
pricing.*USD per 1M tokens
context_windowToken count
capabilities.*Boolean flags (true = supported)
modalities.*Array of: "text", "image", "audio", "video"
performance / speed1–5 scale
status"active" | "deprecated" | "preview"

Markdown format

Query param?format=md
Accept headerAccept: text/markdown
Supported paths/, /:provider, /:provider/:model