ListenHub

ListenHub is an API-first content generation platform in the AI Podcast Assistant space. It accepts text input and returns scripted podcast episodes with synthesized AI voices, positioning itself as a programmatic alternative to manual recording workflows. The service operates under the marswave.ai domain, with its production API at https://api.marswave.ai/openapi.

How it works. Developers authenticate with a Bearer token, storing the key as the LISTENHUB_API_KEY environment variable. The API supports multiple generation modes: podcast (1–2 speakers, single or dual host), debate (exactly 2 speakers), and multi-role dialogue — the last enabling audiobook and radio-drama scenarios where multiple distinct voices appear in a single request. Each episode produces two artifacts: a full script text and an MP3 audio file available for download. Outline and script data are streamed via SSE while audio renders in the background, allowing applications to display episode structure before the MP3 finishes.

Speaker management. Available voices are queried through GET /v1/speakers/list, with each speaker identified by a speakerId and tagged with language and gender metadata. This lets developers programmatically select voices that match the tone and demographic requirements of their content.

API design and error handling. All responses return HTTP 200; the body's code field distinguishes success (0) from failure (≠0). Documented error codes cover invalid keys (21007), insufficient credits (26004), invalid episode state (25008), and parameter validation failures (29003). The vendor explicitly warns that API keys are equivalent to account credentials — a leaked key enables unauthorized credit consumption with no secondary authentication barrier.

What's documented vs. claimed. The vendor's tagline promises more output modalities than the OpenAPI documentation details. The documented endpoints focus on podcast and speech generation; music, image, and video capabilities are not represented in the API reference. The .well-known/openapi.json endpoint returns a 404 page, which complicates automated SDK generation and standards-based API discovery.

Pricing and environments. ListenHub uses a credit-based system with subscription plans and rate limits. Credits are consumed per generation request, and users check their balance via GET /v1/user/subscription. Concrete per-request pricing is not disclosed in the public documentation. A sandbox or test environment is available only by contacting support@marswave.ai — there is no self-service provisioning.

Where it fits. Unlike Riverside, which provides a recording studio for human podcasters, ListenHub automates the entire production chain — script writing, voice synthesis, and audio rendering — through an API. This trades the editorial precision of human recording for throughput and automation. Compared to PodLM, another AI podcast generator, ListenHub claims broader output types but does not yet document them at the level a developer would need to integrate them reliably. For teams building content pipelines, news audio summaries, or automated narration at scale, the API-first design and SSE streaming are structural advantages over GUI-bound alternatives. The absence of a self-service sandbox and the gap between marketing claims and documented endpoints are, however, friction points for technical evaluation.