> ## Documentation Index
> Fetch the complete documentation index at: https://docs.opensink.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Examples

> Working example agents you can run and learn from.

## Open-source examples

These are complete, runnable agents built with OpenSink. Each one demonstrates a different set of primitives and patterns.

All examples are available in the [opensink-examples](https://github.com/opensinkai/opensink-examples) repository.

***

## Finance News Agent

A financial news agent that fetches articles, uses AI to analyze and categorize them, and stores the results in an OpenSink sink.

**What it demonstrates:**

* Sessions for tracking execution
* Configurations for controlling agent behavior (sources, limits)
* Activities for logging what the agent did
* Sink Items for storing curated results

**Stack:** Fastify, OpenAI, NewsAPI

<Card title="Finance News Agent" icon="newspaper" href="https://github.com/opensinkai/opensink-examples/tree/main/examples/1-finance-news-agent-fastify">
  View on GitHub
</Card>

***

## Trading Agent

A trading agent with a human-in-the-loop approval workflow. Analyzes market news, proposes trades, waits for human approval, then executes.

**What it demonstrates:**

* Sessions with durable state for tracking proposed trades
* Input Requests for pausing and asking for human approval
* Execution Endpoints for automatic resumption after approval
* Sink Items for recording executed trades

**Stack:** Fastify, OpenAI, NewsAPI

<Card title="Trading Agent" icon="chart-line" href="https://github.com/opensinkai/opensink-examples/tree/main/examples/2-trading-agent-fastify">
  View on GitHub
</Card>