starred/anything-llm

Fork 0

mirror of https://github.com/Mintplex-Labs/anything-llm.git synced 2026-03-02 22:57:05 -05:00

[FEAT]: Allow users to add transcription of a meeting in the Meeting Assistant over being forced to use audio source #3128

New issue

Open

opened 2026-02-28 06:31:08 -05:00 by deekerman · 2 comments

deekerman commented

2026-02-28 06:31:08 -05:00

Owner

Originally created by @GabrieleGrezzana on GitHub (Jan 29, 2026).

What would you like to see?

This could be handy for those situations where meetings are in languages not yet supported by the Transcription Model and user is forced to process speech-to-text outside AnythingLLM.

Current workaround is to create a new chat, using the transcription for the LLM to create minutes/summaries and interacting with the text.

Proposing this feature as with v1.10.0 we now have this amazing Meeting Assistant feature.

Originally created by @GabrieleGrezzana on GitHub (Jan 29, 2026). ### What would you like to see? This could be handy for those situations where meetings are in languages not yet supported by the Transcription Model and user is forced to process speech-to-text outside AnythingLLM. Current workaround is to create a new chat, using the transcription for the LLM to create minutes/summaries and interacting with the text. Proposing this feature as with v1.10.0 we now have this amazing Meeting Assistant feature.

deekerman added the

labels

2026-02-28 06:31:08 -05:00

deekerman commented

2026-02-28 06:31:11 -05:00

Author

Owner

@timothycarambat commented on GitHub (Jan 29, 2026):

Do you happen to have an example of what these text transcriptions are formatted like? Is it just one big chunk of text or is it JSON? Curiously, where are you getting transcripts from, and ultimately what would you like do be done with them?

I presume skipping the transcript, but doing the summary, agent items, and such?

@timothycarambat commented on GitHub (Jan 29, 2026): Do you happen to have an example of what these text transcriptions are formatted like? Is it just one big chunk of text or is it JSON? Curiously, where are you getting transcripts from, and ultimately what would you like do be done with them? I presume skipping the transcript, but doing the summary, agent items, and such?

deekerman commented

2026-02-28 06:31:13 -05:00

Author

Owner

@GabrieleGrezzana commented on GitHub (Jan 30, 2026):

Do you happen to have an example of what these text transcriptions are formatted like? Is it just one big chunk of text or is it JSON? Curiously, where are you getting transcripts from, and ultimately what would you like do be done with them?

It is pure text with punctuation coming from Vibe, which supports Chinese and Japanese. There is also the option to have Speaker identification, just like with AnythingLLM via Parakeet 0.6B v3.

I presume skipping the transcript, but doing the summary, agent items, and such?

Summary, with the current version of AnythingLLM, can be performed only if the Meeting Assistant successfully transcribe the audio. Letting the users to provide a transcription rather than audio only, would allow them to create summaries, use agents and chat.

Can I summarize the meeting directly via a regular chat using the transcript I have? Sure
Would it be nice to have it through the Meeting Agent? Yes ;)

@GabrieleGrezzana commented on GitHub (Jan 30, 2026): > Do you happen to have an example of what these text transcriptions are formatted like? Is it just one big chunk of text or is it JSON? Curiously, where are you getting transcripts from, and ultimately what would you like do be done with them? > It is pure text with punctuation coming from [Vibe](https://github.com/thewh1teagle/vibe), which supports Chinese and Japanese. There is also the option to have Speaker identification, just like with AnythingLLM via [Parakeet 0.6B v3](https://huggingface.co/nvidia/parakeet-tdt-0.6b-v3). > I presume skipping the transcript, but doing the summary, agent items, and such? Summary, with the current version of AnythingLLM, can be performed only if the Meeting Assistant successfully transcribe the audio. Letting the users to provide a transcription rather than audio only, would allow them to create summaries, use agents and chat. Can I summarize the meeting directly via a regular chat using the transcript I have? Sure Would it be nice to have it through the Meeting Agent? Yes ;)

No milestone

No project

No assignees

1 participant

Notifications

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference

starred/anything-llm#3128

No description provided.

Rows
Columns