TypeWhisper
DEGitHub
Back to Add-ons

OpenAI Compatible

Bundled

by TypeWhisper

TranscriptionLLM macOSWindows
macOS Source Code Windows Source Code
OpenAI Compatible settings

About

OpenAI Compatible is a universal adapter that connects TypeWhisper to any service with an OpenAI-compatible API. Use it with Ollama, LM Studio, vLLM, or any other compatible server for local LLM prompt processing. Servers that also implement OpenAI-compatible audio endpoints can be used for transcription. Models are loaded dynamically from the server’s /v1/models endpoint or can be entered manually.

Features

  • Works with any OpenAI-compatible API (Ollama, LM Studio, vLLM, etc.)
  • Transcription and LLM support
  • Translation support
  • 99+ languages (depending on the model)
  • Dynamic model discovery via /v1/models endpoint
  • Manual model entry supported
  • Test connection function
  • No vendor lock-in

Transcription & LLM Models

Models are loaded dynamically from the connected server. You can also enter model names manually.

Configuration

  • Server URL - The base URL of your OpenAI-compatible server (e.g. http://localhost:11434/v1 for Ollama)
  • API Key - Optional, depends on the server

Quick Setup

  1. Start your OpenAI-compatible server (e.g. Ollama, LM Studio)
  2. Open TypeWhisper Settings > Plugins
  3. Find the OpenAI Compatible plugin and click Configure
  4. Enter your server URL
  5. Use “Test Connection” to verify the connection
  6. Select OpenAI Compatible as your transcription engine or LLM provider

Use Ollama for Local LLM Prompts

Ollama exposes an OpenAI-compatible API on http://localhost:11434/v1. TypeWhisper can use that endpoint for workflows and prompt processing without loading a second copy of the same local LLM.

  1. Install and start Ollama.
  2. Pull a model, for example ollama pull llama3.2.
  3. In TypeWhisper, open Settings > Plugins > OpenAI Compatible.
  4. Set Server URL to http://localhost:11434/v1.
  5. Leave API Key empty.
  6. Click Test Connection.
  7. Refresh models and select the Ollama model you want for LLM tasks.

See the official Ollama OpenAI compatibility docs for the supported /v1/* endpoints.

Use LM Studio for Local LLM Prompts

LM Studio provides an OpenAI-compatible local server, commonly on http://localhost:1234/v1.

  1. Open LM Studio and load a chat model.
  2. Start the local server from LM Studio.
  3. In TypeWhisper, open Settings > Plugins > OpenAI Compatible.
  4. Set Server URL to http://localhost:1234/v1.
  5. Leave API Key empty unless your LM Studio server requires one.
  6. Click Test Connection.
  7. Refresh models and choose the loaded LM Studio model for LLM tasks.

See the official LM Studio OpenAI compatibility docs for the supported endpoints.

Transcription Notes

Ollama and LM Studio are LLM servers. They support OpenAI-compatible text endpoints such as /v1/models and /v1/chat/completions, which is what TypeWhisper uses for prompt processing.

To use OpenAI Compatible as a transcription engine, the server must also implement OpenAI-compatible audio endpoints such as /v1/audio/transcriptions. If your server only supports chat completions, use OpenAI Compatible for LLM workflows and keep a dedicated TypeWhisper transcription engine selected for speech-to-text.

Troubleshooting

  • If Test Connection fails, confirm that the server is running and that the URL ends in /v1.
  • If no models appear, enter the model name manually exactly as the server reports it.
  • If LLM workflows work but transcription fails, the server likely does not implement OpenAI-compatible audio transcription endpoints.