Tone, certainty, detail, and timing, for every line on every podcast

We just shipped a feature our customers have been asking about for a long time. The Particle API now labels every line spoken on a podcast by its tone, how sure the speaker sounds, how specific they get, and whether they’re talking about the past or the future. We call it Speaker Statement Characteristics, and you can search and filter by all four:

Here's what that looks like on podcasts from the last few months.

1. Confident predictions about interest rates

Guests on podcasts, talking about Fed rate cuts, sounding sure about what happens next.

curl -G "https://api.particle.pro/v1/podcasts/episodes/search" \

  -H "X-API-Key: $PARTICLE_API_KEY" \

  --data-urlencode "semantic_search=Federal Reserve interest rate cuts" \

  --data-urlencode "since=2026-05-05" \

  --data-urlencode "speaker_role=guest" \

  --data-urlencode "speaker_statement_characteristics[certainty]=confident" \

  --data-urlencode "speaker_statement_characteristics[temporal_orientation]=future"

• "Lower energy prices will deliver lower inflation as night follows day." Neil Woodford, Noise Cancelling with Neil Woodford · ► Listen

• "And when we start cutting rates, all these stocks are going to explode." George Seay, Street Talk · ► Listen

• "If inflation's at two and base rate's at three, I can absolutely guarantee gilt yields will not be at 5.6%." Neil Woodford, Noise Cancelling with Neil Woodford ·
  ► Listen

2. ...and the people who aren't so sure

Ask for hedged lines instead, and you get the economists keeping their options open:

curl -G "https://api.particle.pro/v1/podcasts/episodes/search" \

-H "X-API-Key: $PARTICLE_API_KEY" \

--data-urlencode "semantic_search=will there be a recession" \

--data-urlencode "since=2026-05-05" \

--data-urlencode "speaker_statement_characteristics[certainty]=hedged"

• "The debate is whether it's going to be a slowdown to 1.5% growth over the next 12 months and whether it will go below that and maybe risk recession." Shane Oliver, Wealth Management Podcast Network · ► Listen

3. Bold, specific calls about AI and jobs

Guests making specific (numbers, dates, named groups), confident, forward-looking claims about AI and work.

curl -G "https://api.particle.pro/v1/podcasts/episodes/search" \

-H "X-API-Key: $PARTICLE_API_KEY" \

--data-urlencode "semantic_search=AI replacing white collar jobs" \

--data-urlencode "since=2026-05-05" \

--data-urlencode "speaker_role=guest" \

--data-urlencode "speaker_statement_characteristics[certainty]=confident" \

--data-urlencode "speaker_statement_characteristics[specificity]=specific" \

--data-urlencode "speaker_statement_characteristics[temporal_orientation]=future"

• "The AI is going to displace 100 million jobs in the next four years." Kerry Kennedy, Next Question with Katie Couric · ► Listen

• "In the next 24 months, I can tell you that jobs involving repetitive, predictable, computer-based tasks face the greatest immediate risk." Brad Aman, Build with Sophia · ► Listen

Leave out the search text entirely and scope it to one person or episode, and you get their strongest lines of that kind, ranked.

curl -G "https://api.particle.pro/v1/podcasts/episodes/search" \

-H "X-API-Key: $PARTICLE_API_KEY" \

--data-urlencode "episode_id=arvind-krishna-nikesh-arora-quantum-threat-open-vs-frontier-ai-ai-cyber-warfare" \

--data-urlencode "speaker_statement_characteristics[specificity]=specific" \

--data-urlencode "speaker_statement_characteristics[temporal_orientation]=future"

• "The debates are is it 2028, 2029, 2033? And sort of if you pick the middle point, let's call it 2030, well, it will break today's cryptography." Arvind Krishna, All-In ·
  ► Listen

• "On data centers, I think there's no chance all 50 will ban." Arvind Krishna, All-In · ► Listen

4. What the experts predict, combined with Speaker Expertise

Statement characteristics work with our Speaker Expertise filters, so you can ask what a whole profession is saying about the future. Cardiologists, talking about GLP-1 drugs, looking ahead:

curl -G "https://api.particle.pro/v1/podcasts/episodes/search" \

-H "X-API-Key: $PARTICLE_API_KEY" \

--data-urlencode "semantic_search=GLP-1 drugs like Ozempic" \

--data-urlencode "since=2026-05-05" \

--data-urlencode "guest_occupation=cardiologists" \

--data-urlencode "speaker_statement_characteristics[specificity]=specific" \

--data-urlencode "speaker_statement_characteristics[temporal_orientation]=future"

• "So presumably over the next few months, some of the generic manufacturers are going to start making generic versions of Ozempic." Dr. Christopher Labos, Montreal Now with Aaron Rand · ► Listen

• "Today, September 30th, 2026, this is the worst the GLP-1 drugs are going to be as we move forward." Dr. Greg Katz, Future Pulse Cardiology · ► Listen

5. How a whole show talks

Every show now has a profile, with percentiles against every other profiled podcast. It’s useful for picking which shows to pitch, sponsor, or monitor, and for knowing in advance whether a show trades in hot takes or careful caveats.

6. How your company is being talked about

Point it at a brand and track the tone over time:

• Share of voice comparison: OpenAI, Anthropic, and Google side by side. Since August, Google drew both the most negative (4.9%) and most positive (5.5%) mentions of the three, across 385,000 mentions.

• Timeseries: "Is Tesla spoken of more negatively this quarter?" is one call, broken down by month.

• Alerts: get pinged only when your company is mentioned negatively, or only when someone makes a prediction about it, by email, Slack, or webhook.

• Your company's people: how your executives come across on podcasts, week by week or month by month.

Questions, or want help with a specific use case? Just reply to this email or write to api@particle.pro.