Skip to main content
Every transcribed episode has two transcript representations:
  • Dialogue — speaker-attributed lines with sentence-level timestamps (fast to retrieve)
  • Word-level — every word individually timestamped with speaker attribution (larger payload)
Both are accessible for full episodes, individual segments, and individual clips.

​
Dialogue transcript

The default format. Each line is a speaker turn with start/end timestamps. 
curl "https://api.particle.pro/v1/podcasts/episodes/6OdMXSZK8XxM89hbS3eanr/transcript" \
  -H "X-API-Key: YOUR_API_KEY"
​
episode_id
string
required
The parent episode ID.
​
language
string
ISO 639-1 language code.
​
duration_seconds
number
required
Total episode duration.
​
lines
DialogueLine[]
required
Example response
{
  "episode_id": "6OdMXSZK8XxM89hbS3eanr",
  "language": "en",
  "duration_seconds": 4832.5,
  "lines": [
    {
      "number": 1,
      "speaker": "Kevin Roose",
      "role": "HOST",
      "start_seconds": 0.0,
      "end_seconds": 4.2,
      "text": "Welcome back to Hard Fork. I'm Kevin Roose."
    },
    {
      "number": 2,
      "speaker": "Casey Newton",
      "role": "CO_HOST",
      "start_seconds": 4.5,
      "end_seconds": 8.1,
      "text": "And I'm Casey Newton. Today we're talking about AI regulation."
    }
  ]
}

​
Output formats

Use the format query parameter. All three formats support speaker filtering and time-range clipping.
The default. Structured JSON with speaker attribution, roles, and timestamps per line.
curl ".../transcript?format=dialogue"
Best for: building conversation UIs, speaker analysis, programmatic processing.

​
Filter by speaker

Extract everything a specific person said: 
curl ".../transcript?speaker=Casey+Newton" \
  -H "X-API-Key: YOUR_API_KEY"
Only lines from that speaker are returned. Works with all three output formats.

​
Time-range clipping

Extract a portion of the transcript by time range: 
# Get the transcript from 5 minutes to 15 minutes
curl ".../transcript?start=300&end=900" \
  -H "X-API-Key: YOUR_API_KEY"
Combine with speaker filtering to get exactly what one person said during a specific discussion: 
curl ".../transcript?speaker=Casey+Newton&start=300&end=900&format=text"

​
Word-level transcript

For applications that need per-word timing — karaoke-style highlighting, precision audio editing, or word-aligned search: 
curl "https://api.particle.pro/v1/podcasts/episodes/6OdMXSZK8XxM89hbS3eanr/transcript/words?start=0&end=10" \
  -H "X-API-Key: YOUR_API_KEY"
Example response
{
  "episode_id": "6OdMXSZK8XxM89hbS3eanr",
  "language": "en",
  "words": [
    {"text": "Welcome", "type": "word", "start_seconds": 0.0, "end_seconds": 0.4, "speaker": "Kevin Roose"},
    {"text": " ", "type": "spacing", "start_seconds": 0.4, "end_seconds": 0.45},
    {"text": "back", "type": "word", "start_seconds": 0.45, "end_seconds": 0.7, "speaker": "Kevin Roose"},
    {"text": " ", "type": "spacing", "start_seconds": 0.7, "end_seconds": 0.75},
    {"text": "to", "type": "word", "start_seconds": 0.75, "end_seconds": 0.9, "speaker": "Kevin Roose"}
  ]
}
Word-level transcripts for long episodes can be very large. Always use start and end parameters to request only the time range you need.

​
Segment and clip transcripts

Segments and clips have their own transcript endpoints that return the dialogue scoped to that time range: 
# Segment transcript
curl "https://api.particle.pro/v1/podcasts/segments/9fLwNa2mK7Rp4xYb/transcript" \
  -H "X-API-Key: YOUR_API_KEY"

# Clip transcript
curl "https://api.particle.pro/v1/podcasts/clips/3hR8jUVn5cQwDzXe/transcript?format=srt" \
  -H "X-API-Key: YOUR_API_KEY"
These support the same format parameter as the episode transcript.