Create Podcast
Create and auto-convert a podcast project. Currently, the LLM cost is covered by us but you will still be charged for the audio generation. In the future, you will be charged for both the LLM and audio generation costs.
Headers
xi-api-key
Request
This endpoint expects an object.
model_id
The ID of the model to be used for this Studio project, you can query GET /v1/models to list all available models.
mode
The type of podcast to generate. Can be 'conversation', an interaction between two voices, or 'bulletin', a monologue.
source
The source content for the Podcast.
quality_preset
Output quality of the generated audio. Must be one of:
standard - standard output format, 128kbps with 44.1kHz sample rate.
high - high quality output format, 192kbps with 44.1kHz sample rate and major improvements on our side. Using this setting increases the credit cost by 20%.
ultra - ultra quality output format, 192kbps with 44.1kHz sample rate and highest improvements on our side. Using this setting increases the credit cost by 50%.
ultra lossless - ultra quality output format, 705.6kbps with 44.1kHz sample rate and highest improvements on our side in a fully lossless format. Using this setting increases the credit cost by 100%.
Allowed values:
duration_scale
Duration of the generated podcast. Must be one of: short - produces podcasts shorter than 3 minutes. default - produces podcasts roughly between 3-7 minutes. long - produces podcasts longer than 7 minutes.
Allowed values:
language
An optional language of the Studio project. Two-letter language code (ISO 639-1).
intro
The intro text that will always be added to the beginning of the podcast.
outro
The outro text that will always be added to the end of the podcast.
instructions_prompt
Additional instructions prompt for the podcast generation used to adjust the podcast's style and tone.
highlights
A brief summary or highlights of the Studio project's content, providing key points or themes. This should be between 10 and 70 characters.
callback_url
A url that will be called by our service when the Studio project is converted. Request will contain a json blob containing the status of the conversion
Messages:
1. When project was converted successfully:
{
type: "project_conversion_status",
event_timestamp: 1234567890,
data: {
request_id: "1234567890",
project_id: "21m00Tcm4TlvDq8ikWAM",
conversion_status: "success",
project_snapshot_id: "22m00Tcm4TlvDq8ikMAT",
error_details: None,
}
}
2. When project conversion failed:
{
type: "project_conversion_status",
event_timestamp: 1234567890,
data: {
request_id: "1234567890",
project_id: "21m00Tcm4TlvDq8ikWAM",
conversion_status: "error",
project_snapshot_id: None,
error_details: "Error details if conversion failed"
}
}
3. When chapter was converted successfully:
{
type: "chapter_conversion_status",
event_timestamp: 1234567890,
data: {
request_id: "1234567890",
project_id: "21m00Tcm4TlvDq8ikWAM",
chapter_id: "22m00Tcm4TlvDq8ikMAT",
conversion_status: "success",
chapter_snapshot_id: "23m00Tcm4TlvDq8ikMAV",
error_details: None,
}
}
4. When chapter conversion failed:
{
type: "chapter_conversion_status",
event_timestamp: 1234567890,
data: {
request_id: "1234567890",
project_id: "21m00Tcm4TlvDq8ikWAM",
chapter_id: "22m00Tcm4TlvDq8ikMAT",
conversion_status: "error",
chapter_snapshot_id: None,
error_details: "Error details if conversion failed"
}
}
apply_text_normalization
This parameter controls text normalization with four modes: 'auto', 'on', 'apply_english' and 'off'.
When set to 'auto', the system will automatically decide whether to apply text normalization
(e.g., spelling out numbers). With 'on', text normalization will always be applied, while
with 'off', it will be skipped. 'apply_english' is the same as 'on' but will assume that text is in English.
Allowed values:
Response
Successful Response
project
The project associated with the created podcast.
