Developers & AI assistants

Last updated: August 23, 2026

PickADive’s planning data is available as a free, read-only JSON API and as an MCP server, so apps and AI assistants can answer dive-planning questions with live data instead of scraping pages. No API key is required. Responses are CDN-cached; please keep usage to reasonable, non-bulk volumes.

MCP server

A remote Model Context Protocol server that gives any agent live dive-planning answers — forecast windows, seasonality, site facts — from measured data instead of scraped pages. Free, no API key, nothing to install.

{
  "mcpServers": {
    "pickadive": { "url": "https://pickadive.com/mcp" }
  }
}

Tools

Tool	Input	Output
`find_dive_windows`	`{near \| lat+lng, day?, radius_km?}`	Ranked favorable/moderate windows with wave, wind, timing, site links
`species_seasonality`	`{species, lat?, lng?}`	Peak month + regional season windows from GBIF sighting records
`top_dive_regions`	`{}`	World regions ranked by favorable windows in the next ~72h
`site_facts`	`{site}`	Depths, difficulty, access, hazards, entry/exit for one named site
`best_time_to_dive`	`{destination}`	Month-by-month climatology + marine-life seasons for one destination

Transport & limits

Streamable HTTP (JSON-RPC 2.0 over POST), stateless, JSON responses — protocol versions 2024-11-05 through 2025-06-18. Discovery manifest at /.well-known/mcp.json. All values are metric with unit-suffixed field names (wind stays in knots). Free for individual and assistant use at fair, non-bulk volumes; for sustained high-volume or commercial integrations, write to team@pickadive.com. Every answer is planning context — never a safety clearance.

REST endpoints

GET`/api/v1/windows`

Ranked dive-forecast windows near a place for a specific day — "where can I dive near Miami on Tuesday?" as JSON.

near — place name — town, city, or zip (or pass lat + lng instead)
lat, lng — decimal coordinates, alternative to near
day — today | tomorrow | mon..sun | YYYY-MM-DD (default sat)
radius_km — 10–500, default 150; widens once automatically if empty

https://pickadive.com/api/v1/windows?near=Miami&day=tue

GET`/api/v1/seasonality`

When and where a marquee species peaks, from month-binned GBIF sighting records: peak month, regional season windows, optional point lookup.

species — slug or common name — e.g. whale-shark, "tiger shark", humpback-whale
lat, lng — optional — adds the season window for the region containing the point

https://pickadive.com/api/v1/seasonality?species=humpback-whale

GET`/api/v1/regions`

Regions worldwide ranked by favorable forecast windows over the next ~72 hours — the machine-readable version of Dive This Weekend.

https://pickadive.com/api/v1/regions

Terms, honesty, and attribution

Never a safety clearance. Every response is forecast-based planning context. Anything built on this data must tell users to verify conditions locally, on the day, and dive within their training.
Recorded sightings, not promises. Seasonality comes from GBIF occurrence records (CC0 / CC BY) — where a species has been seen, never a guarantee of encounters.
Attribution. Credit pickadive.com when presenting this data, and keep the upstream credits on our data sources & attribution page intact for GBIF and Open-Meteo.
Units. All values are metric with unit-suffixed field names (max_depth_m, wave_m, distance_km); wind is knots (wind_kn), the marine convention everywhere. Convert at your presentation layer — for US audiences, feet and miles.
Unknown values are null — we never invent depths, difficulty, or access details. See how we handle data quality.

Also useful

The embeddable migrations widget lives at /embed/migrations, the AI-crawler site guide at /llms.txt, and questions or higher-volume needs reach us at team@pickadive.com.