Core Primitive

Sinks & Items

Sinks are containers. Items are the data. Push anything from your agents with a single API call — no schema setup required.

finance-news

Market alerts & signals

AAPL dropped 3.2%

alert2 min ago
{"ticker":"AAPL","change":-3.2}

NVDA hits all-time high

signal8 min ago
{"ticker":"NVDA","price":892.5}

Fed announces rate decision

news15 min ago
{"impact":"high"}

Create Items in Seconds

Use our SDK or hit the API directly. It's that simple.

create-item.ts
import { OpenSink } from "@opensink/sdk";
const client = new OpenSink({
apiKey: process.env.OPENSINK_API_KEY
});
// Create an item in your sink
await client.items.create("finance-news", {
title: "AAPL dropped 3%",
body: "Apple stock fell following earnings report",
type: "alert",
url: "https://example.com/news/aapl",
fields: {
ticker: "AAPL",
change: -3,
source: "bloomberg"
}
});

Item Schema

titlestringrequired
Item headline
bodystring
Detailed content
typestring
Category label
urlstring
Related link
fieldsobject
Custom key-value data
occurred_atstring
Event timestamp

Built for Agent Data

API Key Protected

Each sink has its own secret key for secure writes

Public or Private

Control visibility with a simple toggle

Searchable

Full-text search across all items in a sink

Flexible Fields

Store any structured data with custom fields

Query & Search

Retrieve items with filters, pagination, and full-text search.

query-items.ts
// List items from a sink
const items = await client.items.list("finance-news", {
limit: 20,
type: "alert"
});
// Get a specific item
const item = await client.items.get("finance-news", "item_123");
// Search items
const results = await client.items.search("finance-news", {
query: "AAPL",
limit: 10
});

Common Use Cases

1,247 items
finance-news

Market updates and stock alerts

892 items
customer-research

User feedback and insights

3,421 items
deployment-logs

CI/CD pipeline results

567 items
agent-findings

AI-discovered information

2,103 items
support-tickets

Customer support events

445 items
trading-signals

Buy/sell recommendations

Why Not Just Use a Database?

Traditional Database

  • Design schema upfront
  • Set up access control
  • Build API layer
  • Handle migrations

OpenSink

  • No schema required
  • Built-in API keys
  • Ready-to-use REST API
  • Just push data

Start pushing data in 2 minutes