{"id":19531426,"url":"https://github.com/ranpy13/stt-meeting-summary","last_synced_at":"2025-10-04T07:45:59.818Z","repository":{"id":224527292,"uuid":"733574451","full_name":"ranpy13/stt-meeting-summary","owner":"ranpy13","description":"A simple machine learning model to translate audio from a meeting to text and then summarize the entire meeting into concise bullet points.  Then the entire package is wrapped into a website, which mails the summarized text file to the administrator and attached email addresses.","archived":false,"fork":false,"pushed_at":"2024-01-05T03:10:23.000Z","size":30614,"stargazers_count":5,"open_issues_count":0,"forks_count":0,"subscribers_count":1,"default_branch":"main","last_synced_at":"2025-03-29T06:22:39.026Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"gpl-3.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/ranpy13.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2023-12-19T16:26:35.000Z","updated_at":"2024-09-23T19:37:24.000Z","dependencies_parsed_at":null,"dependency_job_id":"3de8cb1f-75d5-4d13-ac96-888a8e2751fa","html_url":"https://github.com/ranpy13/stt-meeting-summary","commit_stats":null,"previous_names":["ranpy13/stt-meeting-summary"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ranpy13%2Fstt-meeting-summary","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ranpy13%2Fstt-meeting-summary/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ranpy13%2Fstt-meeting-summary/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ranpy13%2Fstt-meeting-summary/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/ranpy13","download_url":"https://codeload.github.com/ranpy13/stt-meeting-summary/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":249508723,"owners_count":21283456,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2024-11-11T01:42:49.599Z","updated_at":"2025-10-04T07:45:54.780Z","avatar_url":"https://github.com/ranpy13.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Speech to Text | Meeting summarizer model\n---\n## Description\n* The goal is to make a model which **translates audio from meetings and encodes it to a text file**, which on completion is *summarized* in to a smaller, concise *list of bullet points* covering the entire meeting. The initial idea is to implement it through pre-trained models, but the accuracy and efficiency has to take care of.\n* There are a few considersation. First, that it should desirably also support *mulit-lingual conversation*, with **english** being the primary support, and additional support for **hindi** and at least the default regional language.\n* The end goal however, is to wrap this model into a presentable *web-app*, with gui interface to start the recording, and add the administrators to the meetings, which will *recieve the summary as a text file in the mail box* automatically, without manual intervention.\n\n## Problem Statement\n\u003e `To create a machine learning model which can listen to the audio from meetings and translate the speech to text, and finally output a summary of the entire meeting in a text format. This model can further be wrapped inside a graphical interface for easier access, where the summarized text has to be sent to the administrators of the meeting provided on onset.`\n\n## Course of Action\n### Setting the baseline\n* using google's standard api to convert speech to text[^1]\n  * setting a baseline \n  * finding accuracy\n  * improving on the same\n* dumping the output to a text file\n  * that is then picked up for summarizing\n  * using ~~genism~~[^2] [^3] *sumy* standard module[^4] [^5]\n  * improving on the same\n\n### Improving on the baseline\n* having got the baseline in \n  * google's `speechrecognition`[^6] for converting speech to text\n  * `sumy-lsasummarize`[^7] for summarizing the contents\n* to use them together and streamline the model\n  * bridge the gap between the two - **preprocessing**\n  * *process* the output from converted text with proper punctuation and markings\n  * then running the summarizing models on the processed text block\n* thus, require a middleware leveragin a *natural languge processing* model\n  * current options - `nltk`[^8], `openai`[^9], `spacy`[^10]\n* _**major shift** in workflow_, found better option: [^11]\n  * **`vosk` model** for speech to text convertion\n  * preprocessing through **`transforerms` model**\n  * finally, summarizing through _`pipeline()`_\n\n### Advanced Optimization and Improvization\n* Context aware summary\n  * change of paragraph in cases of change of paragraph\n  * might require deep learning,\n  * can seperate two summaries, general and context based summary\n* Multilingual speech optimization\n* Adapting to bandwidth, backup solutions\n  * recording set audio \n  * fall-back to recording and translating the saved audio instead of live time transcribing\n\n\\\n\\\n\u0026nbsp;\u0026nbsp;\u0026nbsp;\u0026nbsp;\n---\n## References \u0026 Bibliography\n[^1]: GfG articles on speech-to-text [Python: Con...](https://www.geeksforgeeks.org/python-convert-speech-to-text-and-text-to-speech/)\n\n[^2]: Knowledge Base/ [turing.com](https://www.turing.com/kb/5-powerful-text-summarization-techniques-in-python)\n\n[^3]: Developer's documentation for [`genism`](https://radimrehurek.com/gensim/)\n\n[^4]: Sasha Bondar's blog post on [reintech.io](https://reintech.io/blog/how-to-create-a-text-summarization-tool-with-sumy-tutorial-for-developers)\n\n[^5]: Official PyPi documentation for [`sumy`](https://pypi.org/project/sumy/) \n\n[^6]: Official PyPi documentation for [`SpeechRecognition`](https://pypi.org/project/SpeechRecognition/)\n\n[^7]: Official PyPi documentation for [`sumy`](https://pypi.org/project/sumy/)\n\n[^8]: Official PyPi documentation for [`nltk`](https://pypi.org/project/nltk/)\n\n[^9]: Official PyPi documentation for [`openai`](https://pypi.org/project/openai/)\n\n[^10]: Official PyPi documentation for [`spacy`](https://pypi.org/project/spacy/)\n\n[^11]: DataQuest's blog post, [github](https://github.com/dataquestio/project-walkthroughs/blob/master/speech_recognition/README.md)","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Franpy13%2Fstt-meeting-summary","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Franpy13%2Fstt-meeting-summary","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Franpy13%2Fstt-meeting-summary/lists"}