Sentiment Workflows

OpenClaw is a good fit for market-intelligence and research loops where you log text, score it, then emit aggregate signals to Chirpier.

Example: monitor Bitcoin sentiment on Reddit

For this workflow, an OpenClaw task typically:

1. searches Reddit for posts matching bitcoin OR btc
2. scores posts or comments with a sentiment model
3. emits aggregate sentiment events every few minutes
4. charts the trend in Chirpier and alerts on negative shifts

For Bitcoin sentiment on Reddit, useful events are:

• sentiment.bitcoin.reddit.score
• sentiment.bitcoin.reddit.positive.count
• sentiment.bitcoin.reddit.negative.count
• sentiment.bitcoin.reddit.post_volume.count

Example payload:

{
  "agent": "openclaw.sentiment-watcher",
  "event": "sentiment.bitcoin.reddit.score",
  "value": -0.62,
  "meta": {
    "source": "reddit",
    "query": "bitcoin OR btc"
  }
}

You can also emit the negative post count in the same run:

{
  "agent": "openclaw.sentiment-watcher",
  "event": "sentiment.bitcoin.reddit.negative.count",
  "value": 143,
  "meta": {
    "source": "reddit"
  }
}

Starter policies:

• sentiment.bitcoin.reddit.score with average < -0.4 in hour
• sentiment.bitcoin.reddit.negative.count with sum > 100 in hour
• sentiment.bitcoin.reddit.post_volume.count with sum > 500 in day

Emit aggregate scores and counts, not raw text. Keep the source and query in meta.

Useful charts for this workflow are:

• average Reddit sentiment by hour
• negative post count by hour
• post volume by day