This is a fork of the Owncast project which aims to implement an online transcription service into Owncast.
As part of my bachelor thesis, I cannot accept pull requests at this point.
Also, please do not expect the master branch to be as reliable as on the Owncast master
branch.
This fork appeared after the v0.0.6 release.
For any issues besides the transcription, please visit the Owncast issues page and our RocketChat.
- Create an AWS account if you don't have one already.
- In IAM, create an user with "Programmatic Access" and access to
AmazonTranscribeFullAccess
. - Generate a key pair and save it in
$HOME/.aws/credentials
:
[default]
aws_access_key_id = <Access key ID>
aws_secret_access_key = <Secret access key>
- Configure the application acc. to the instructions below.
- Start the application with
./start.sh
.
- Ensure you have the gcc compiler configured.
- Install the Go toolchain.
- Install the Microsoft Azure Speech SDK for Go
- Clone the repo.
git clone https://github.com/owncast/owncast
- Update
start.sh
to contain the right paths. ./start.sh
will run from source.- Point your broadcasting software at your new server and start streaming.
- Make sure you have a GCP account with a project and a service account. Activate the "Speech-to-Text" feature and download a JSON key.
- Clone the repo (see Azure).
- Edit
start.sh
to contain the path to your key file. ./start.sh
Copy the transcription-example.yaml
file to transcription.yaml
and edit accordingly.
Edit the configuration according to the documentation.
This build usually uses the WebSockets delivery method, since it works better in most environments.
If you are looking to use WebVTT instead, please edit ./start.sh
to build Owncast with -tags webvtt
.
Distributed under the MIT License. See LICENSE
for more information.
Project chat: Join us on Rocket.Chat if you want to contribute, follow along, or if you have questions.
Jannik Volkland - @[email protected] - email [email protected]
Project Link: https://github.com/owncast/owncast