OpenWhistle

OpenWhistle lets users configure conversational AI agents that process real audio recordings and output structured data in JSON, Markdown, or via webhooks. Users describe an agent in plain English, list the desired output fields, and the system automatically generates the prompt and schema, handling language detection for over twenty languages. Audio can be uploaded, linked, or recorded directly in the browser; long recordings are chunked and processed in parallel without requiring custom code or model management.

The platform is aimed at teams that need to turn calls, interviews, or other voice interactions into actionable information—such as sales qualification, support ticket triage, recruiting, legal intake, and user research. It provides EU‑hosted, encrypted models and supports live recording, replay, and multi‑record handling, with connectors to common messaging services and custom HTTPS endpoints.

OpenWhistle emphasizes a minimal setup: define the agent, capture audio, and receive the structured response, all through a web interface, API, or CLI. It bundles templates for five typical personas and offers a free tier without a credit card, positioning itself as a production‑ready, low‑maintenance solution for turning voice data into structured intelligence.