Open-source platform for 3D AI agents. Turn text or a photo into a rigged, animated GLB avatar, give it an LLM brain, memory and a wallet, and embed it anywhere with one web component. Remote MCP server, x402 payments, Three.js, browser-native.
JavaScript 3d-generationai-agentsanimationavatardigital-humangenerative-aiglbgltfimage-to-3dllmmcpmcp-servermodel-context-protocolsolanatext-to-3dthreejsweb-componentswebglwebxrx402
-- Rolling store of geolocated force, posture and sanctions events for /globe.
--
-- GDELT publishes one 15-minute export file and keeps no queryable history: its
-- GEO 2.0 API has 404'd for every query for long enough that it cannot be built
-- on. So the globe's time control (24h / 7d / 30d) is only honest if we keep the
-- window ourselves. /api/cron/globe-ingest reads each new file as it lands and
-- writes the classified events here; /api/globe/intel reads the window back.
--
-- Only the three CAMEO families the globe renders are stored (conflicts, force
-- posture, sanctions), roughly a tenth of each file, so a 30-day window is tens
-- of thousands of rows rather than millions.
--
-- The primary key is GDELT's own GlobalEventID, which makes re-ingesting an
-- overlapping file a no-op and lets the cron re-read a window it already covered
-- without duplicating anything.
create table if not exists globe_events (
id bigint primary key,
layer text not null check (layer in ('conflicts', 'military', 'sanctions')),
occurred_at timestamptz not null,
lat double precision not null,
lon double precision not null,
place text not null default '',
country text,
cameo_code text not null,
root_code text not null default '',
quad_class smallint,
goldstein real,
mentions integer not null default 1,
avg_tone real,
actor1 text,
actor2 text,
source_url text,
source_stamp text not null,
ingested_at timestamptz not null default now()
);
-- Every read is "this layer, over this window, heaviest first", and the retention
-- sweep is "everything older than this".
create index if not exists globe_events_layer_time_idx on globe_events (layer, occurred_at desc);
create index if not exists globe_events_time_idx on globe_events (occurred_at desc);
-- The cron asks "which files have I already read?" before deciding what to fetch.
create index if not exists globe_events_stamp_idx on globe_events (source_stamp desc);