Request access

Ephemeral into your stack.

v2.4 API

Sub-second latency · 99.97% uptime 3 protocols · REST, Webhooks, MCP

API

Webhooks

MCP

Tested clients

Pull.

Any signal, any window.

Query any topic, any time horizon, any source cut. Sub-second response over REST.

When to use

Polling dashboards

Internal BI tools, terminal overlays, refresh-on-interval views.

One-off queries

Ad-hoc analysis. Notebooks. Quick checks.

Reference

REST Endpoints

GET /

Service metadata (public)

GET /

Service metadata (public)

GET /sources

List all active content sources

GET /sources

List all active content sources

GET /sources/{source_id}

Content items for a source

GET /sources/{source_id}

Content items for a source

GET /sources/item/{item_id}

Transcript for a content item

GET /sources/item/{item_id}

Transcript for a content item

GET /sources/item/{item_id}/full

Transcript + source/item metadata

GET /sources/item/{item_id}/full

Transcript + source/item metadata

GET /content-items/search

Full-text search across content

GET /content-items/search

Full-text search across content

GET /content-items/search-by-ticker

Ticker-based search, grouped by item

GET /content-items/search-by-ticker

Ticker-based search, grouped by item

Push.

When signal breaks.

Define a threshold. Get notified the moment it breaks. Route it to a webhook, an email address, or a push notification. No polling, no missed signals.

When to use

Webhook delivery

POST to any endpoint you control — route it wherever your logic lives.

Email & push alerts

A direct notification the moment a threshold breaks,  on desktop or mobile.

Reference

Trigger types

keyword_match

Fires when a saved keyword is mentioned

speaker_match

Fires when a tracked speaker is mentioned

Delivery targets

Webhook

POST to a client-controlled endpoint)

Sent to one or more addresses

Browser notifications

In-app/browser notification

Agent-native.

No glue code.

Ephemeral ships a native MCP server. Point your agent at it. Your LLM queries real-time signal data directly — no pipelines to maintain, no schema to learn. Ephemeral becomes a tool in your model's toolbox.

When to use

Give agents real-time context

LLMs trained on stale data + Cast for what's happening now.

Skip the data pipeline

No ETL, no warehouse sync, no schema mapping. Just point the agent.

Reference

Trigger types

get_sources

List all active content sources

get_latest_sources

Most recently active sources

get_content_items