Dictation
Audio transcription proxy API behind tenant secrets
Route multipart audio transcription through /dictate using the same tenant-secret boundary as text.
Teams adding voice input or dictation to internal tools without a separate provider credential path.
The problem
Dictation integrations often grow a separate security and provider configuration path from text generation, even when the same apps need both.
How LLM Proxy helps
LLM Proxy exposes POST /dictate for multipart audio and routes dictation-capable providers behind the same tenant-secret authentication model.
How it works
- Send multipart/form-data to /dictate with audio or file as the audio part.
- Authenticate with key=<tenant secret> in the query string.
- Omit provider and model for tenant defaults or select a dictation-capable provider.
- Receive JSON with the transcribed text.
Feature-to-benefit table
| Feature | Why it matters here | Example |
|---|---|---|
| /dictate endpoint | Audio transcription has a public contract separate from text but shares authentication. | The success response is JSON with a text field. |
| Dictation model catalogs | OpenAI, SiliconFlow, Z.AI, and Grok/xAI dictation models are configured explicitly. | Unknown dictation models fail at the edge. |
| Audio size limit | max_input_audio_bytes protects the upstream call boundary. | Oversized audio returns a client error before routing. |
Use-case examples
Voice note ingestion
An internal app uploads recording.webm to /dictate and receives a transcript.
Provider-specific transcription
A caller selects provider=siliconflow when that tenant has a configured key.
Shared governance
Text and voice traffic use one tenant-secret model and shared usage metadata rules.
Objections and limitations
- Only providers with implemented dictation adapters are available through /dictate.
- The audio part is required; missing or invalid multipart forms return 400.
- Not all provider text models are dictation models.
FAQ
What is the main job of audio transcription proxy API?
LLM Proxy exposes POST /dictate for multipart audio and routes dictation-capable providers behind the same tenant-secret authentication model.
Who should read this dictation resource?
Teams adding voice input or dictation to internal tools without a separate provider credential path.
Does this page claim provider performance or pricing advantages?
No. The supported claim is about LLM Proxy's documented routing, configuration, management, security, usage, and deployment contracts. Provider cost, speed, rankings, and benchmark claims are not made here.
Where should setup details come from?
Use the main README and implementation notes for current command, config, and endpoint details. This page summarizes the workflow without replacing LLM Proxy documentation.
What should I read next?
A closely related resource is Dictation provider routing for OpenAI, Z.AI, Grok, and SiliconFlow, which covers dictation provider routing.
Related resources
Use this pattern in LLM Proxy
Start from the canonical API reference, then use the management surface when the workflow needs tenant or provider configuration.
Open API reference