Streaming (Real-time) Transcription
1) Test Streaming accuracy and latency without writing any code:
Go to https://www.rev.ai/Streaming, click on the red microphone, and talk. No login required.
You can also test Streaming at https://www.rev.ai/account/liveStreaming and try out additional options, including:
- Additional languages: French, German, Italian, Japanese, Korean, Mandarin, Portuguese, and Spanish.
- Use of our v2 ASR model. More information here.
- Profanity filter
- Disfluencies removal
- Custom Vocabulary
To test Streaming using these options, register for a free account: https://www.rev.ai/auth/signup and go to https://www.rev.ai/account/liveStreaming.
When finished, you can go to the Speech to Text Jobs dashboard https://www.rev.ai/jobs/speech-to-text and either view or download a txt, json, or srt or vtt caption files.
Note: The first pass is light gray and becomes black as the final word choice is done (demonstrating the context based transcription). Also notice capitalization and punctuation. |
2) To test Streaming via the API:
This testing method requires registering for an account.
- Register for a free account: https://www.rev.ai/auth/signup
- Get your access token: https://www.rev.ai/access_token
- Walk through the Streaming Documentation here: https://docs.rev.ai/api/streaming/get-started/
- Specify the transcription language with the language query parameter. If this is not supplied, transcription will default to English.
- Additional request options for Streaming can be found at https://docs.rev.ai/api/streaming/requests/
Streaming SDK Code Samples can be found here: https://docs.rev.ai/api/streaming/code-samples/
Asynchronous Transcription
These testing methods require registering for an account.
- Register for a free account here: https://www.rev.ai/auth/signup and login
1) To test Asynchronous Transcription (file based batch/pre-recorded) without writing any code:
- Go to the dashboard https://www.rev.ai/jobs/speech-to-text and click the SUBMIT A JOB button.
- Click the UPLOAD FILE button and select a file to transcribe from your computer.
- Specify the source language and click the NEXT button
- Full list of languages offered can be found here
- Optional: you may add Custom Vocabulary for your submission (enter one word or phrase per line)
When finished, you can go to the Speech to Text Jobs dashboardhttps://www.rev.ai/jobs/speech-to-text and either view or download a txt, json, or srt or vtt caption files.
2) To test Asynchronous Transcription via the API:
This testing method requires registering for an account.
- Register for a free account: https://www.rev.ai/auth/signup
- Get your access token: https://www.rev.ai/access_token
- Walk through the Asynchronous Documentation here: https://docs.rev.ai/api/asynchronous/get-started/
- Additional request options for Asynchronous Transcription can be found at https://docs.rev.ai/api/asynchronous/reference/#operation/SubmitTranscriptionJob
Asynchronous SDK Code Samples can be found here: https://docs.rev.ai/api/asynchronous/code-samples/
In all cases, be sure to notice the punctuation and capitalization. You can also try using dates and phone numbers to see how our Inverse Text Normalization works. More information on these features here: https://docs.rev.ai/api/features/#punctuation-and-inverse-text-normalization
Language Identification
Testing Language Identification is only available via the API:
This testing method requires registering for an account.
- Register for a free account: https://www.rev.ai/auth/signup
- Get your access token: https://www.rev.ai/access_token
- Walk through the Language Identification Documentation here: https://docs.rev.ai/api/language-identification/get-started/
- Additional request options for Language Identification can be found at https://docs.rev.ai/api/language-identification/reference/#operation/SubmitLanguageIdentificationJob
Language Identification SDK Code Samples can be found here: https://docs.rev.ai/api/language-identification/code-samples/
Language Identification outputs come in the form of a JSON file with the most probable dominant language from the input audio. You can download the JSON output from any of your Language Identification jobs here.
Example of a Language Identification output:
Topic Extraction
These testing methods require registering for an account.
- Register for a free account here: https://www.rev.ai/auth/signup and login
1) To test Topic Extraction without writing any code:
- Go to the dashboard https://www.rev.ai/jobs/topic-extraction and click the SUBMIT A JOB button.
- Click the UPLOAD FILE button and select a txt or json file from your computer.
- Review the submission details
- Click SUBMIT
- Once your file has been submitted, you’ll see it in the list on the Topic Extraction Jobs dashboard https://www.rev.ai/jobs/topic-extraction and can click on the blue link to view it, or download it as a json file.
- When viewing the Topic Extraction job, you can see the text of the file with a list of topics, including blue highlights on the sentences which contain them. You can use the Score Threshold slider in the upper left to adjust the confidence in the file’s topics, with a higher confidence level showing less topics and vice versa.
2) To test Topic Extraction via the API:
This testing method requires registering for an account.
- Register for a free account: https://www.rev.ai/auth/signup
- Get your access token: https://www.rev.ai/access_token
- Walk through the Topic Extraction Documentation here: https://docs.rev.ai/api/topic-extraction/get-started/
- Additional request options for Topic Extraction jobs can be found at https://docs.rev.ai/api/topic-extraction/reference/#operation/SubmitTopicExtractionJob
Topic Extraction SDK Code Samples can be found here: https://docs.rev.ai/api/topic-extraction/code-samples/
Sentiment Analysis
These testing methods require registering for an account.
- Register for a free account here: https://www.rev.ai/auth/signup and login
1) To test Sentiment Analysis without writing any code:
- Go to the dashboard https://www.rev.ai/jobs/sentiment-analysis and click the SUBMIT A JOB button.
- Click the UPLOAD FILE button and select a txt or json file from your computer.
- Review the submission details
- Click SUBMIT
- Once your file has been submitted, you’ll see it in the list on the Sentiment Analysis Jobs dashboard https://www.rev.ai/jobs/sentiment-analysis and can click on the blue link to view it, or download it as a json file.
- When viewing the Sentiment Analysis job, you can see the text of the file with positive statements highlighted in green, negative statements highlighted in red, and neutral statements without any highlights. When hovered, positive and negative sentence highlights provide a sentiment type, sentiment score (between -1 and 1), and location in the text, denoted by offset and length. In the upper left, you can select between positive and negative sentiment.
2) To test Sentiment Analysis via the API:
This testing method requires registering for an account.
- Register for a free account: https://www.rev.ai/auth/signup
- Get your access token: https://www.rev.ai/access_token
- Walk through the Sentiment Analysis Documentation here: https://docs.rev.ai/api/sentiment-analysis/get-started/
- Additional request options for Sentiment Analysis jobs can be found at https://docs.rev.ai/api/sentiment-analysis/reference/#operation/SubmitSentimentAnalysisJob
Sentiment Analysis SDK Code Samples can be found here: https://docs.rev.ai/api/sentiment-analysis/code-samples/
Additional Resources
- A complete list of languages offered for our products
- Custom Vocabulary Documentation
- Video Demo for Submitting Language Identification, Topic Extraction, and Sentiment Analysis jobs