AirGapScribe
Freemium
Privacy-focused professionals in legal, health, or finance
Whisper
FreeLocal model: free; hosted API: $0.006/audio minute until retirementAudio to textOpen SourceAPISelf-Hosted
Whisper is OpenAI's speech-recognition model family for multilingual transcription and translation into English, available as open-source models and a separately metered hosted API.
Similar tools
Good TapeFreemiumGood Tape is a browser-based transcription service built specifically for professional…
TrainScriptionFreemiumTrainScription runs Whisper entirely in your browser via WebAssembly, processing audio in…
Wispr FlowFreemiumFlow works on a hotkey: hold it, speak, release, and polished text appears wherever your cursor…
DictawizFreemiumThe tool is backed by Google Cloud TTS and surfaces 900+ voices across 50+ languages through a…
Pricing
- Model
- Free local model / Usage-based API
- Price
- Local model: free; hosted API: $0.006/audio minute until retirement
- Free Tier
- The local model is available without a hosted API usage charge, but requires computing resources. No recurring free hosted whisper-1 allowance is established by these sources.
Summary
The local code and model weights use the MIT License. Local transcription requires software setup and computing resources. That local option should not be confused with a free hosted API allowance. OpenAI lists whisper-1 at $0.006 per minute, about $0.36 for an hour of audio. Its API reference identifies the hosted model as based on open-source Whisper V2. OpenAI announced deprecation on August 26, 2026, with API removal scheduled for February 26, 2027. Existing integrations need a migration plan; the notice concerns the hosted service rather than declaring downloaded weights unavailable. Recorded-file uploads have a 25 MB limit, and whisper-1 does not stream transcription responses.
Bottom line: Use local Whisper where you can manage deployment; existing hosted users should migrate before the announced February 26, 2027 API shutdown.
Find your fit
Before you choose Whisper
More options in Transcription / STT. Compare pricing, platforms, and features side by side.
Fast Transcriber
Freemium
Students and educators needing lecture transcripts
Pros
- Multilingual speech transcription
- MIT-licensed code and weights for local deployment
- Hosted API with per-minute pricing
Cons
- Hosted API retirement scheduled for February 2027
- Local deployment requires software and compute resources
- Transcripts need review for unclear speech and names
About
- Platforms
- Local software, API
- API Available
- Yes
- Self-Hosted
- Yes
- Last Updated
- October 17, 2026
Best For
Who it's for
- Developers maintaining existing transcription workflows
- Researchers evaluating speech recognition
- Teams managing local audio processing
What it does well
- Interview and podcast transcripts
- Subtitle drafts
- Translation of recorded speech into English
Integrations
Compare Whisper
Related Workflow Packs
Recommended skills for this tool
Auto-curated by the AIDiveForge recommendation matrix. These skills are predicted to enhance this tool based on category, capability, and domain signals.
Frequently Asked Questions
- Is Whisper free?
- Whisper offers a free option. Hosting, model usage, or optional services may have separate costs; check the pricing details and official terms.
- Is Whisper open source?
- Yes. Whisper is open source — the source repository is at https://github.com/openai/whisper.
- Does Whisper have an API?
- Yes. Whisper exposes a developer API. See the official documentation at https://developers.openai.com/api/docs/models/whisper-1 for details.
- Can I self-host Whisper?
- Yes. Whisper supports self-hosting on your own infrastructure.
- What platforms does Whisper support?
- Whisper is available on: Local software, API.
People also compared
Whisper is OpenAI’s speech-recognition model family for turning recordings into text. The model documentation describes multilingual transcription, translation of speech into English and language identification. It is useful for preparing interview transcripts, podcast text and subtitle drafts, with human review of names, specialist terms and unclear speech before publication.
The open-source model and OpenAI’s hosted API are different ways to use the technology. The API reference identifies whisper-1 as powered by the open-source Whisper V2 model. The official repository releases the code and model weights under the MIT License. Local deployment requires a compatible software environment and computing resources; it is not the same as a free hosted transcription allowance. The hosted whisper-1 price is $0.006 per audio minute, approximately $0.36 for an hour, before any applicable taxes.
Availability matters for new integrations. OpenAI announced deprecation on August 26, 2026 and plans to remove whisper-1 from its API on February 26, 2027. This scheduled hosted-service retirement should not be confused with a statement that downloaded open-source model files disappear. Existing API users should plan migration, and new projects should compare the currently recommended transcription options.
The speech-to-text guide describes recorded-file uploads with a 25 MB limit and distinguishes transcription in the source language from translation into English. Supported output formats depend on the model, and whisper-1 does not support streamed transcription responses. Check the current API guide for file handling and replacement models before building a long-lived workflow.
Contribute
- Work on Whisper? Claim this listing.
- Embed a badge for your README or site.
Sign in to rate this tool or suggest an edit.