Skip to content
@bitcointranscripts

Bitcoin Transcripts

A historical archive for knowledge preservation and propagation

Bitcoin Transcripts

At Bitcoin Transcripts we are building a largely autonomous transcription workflow that streamlines the AI-generation, review by humans, and publication of bitcoin tech transcripts.

workflow

Transcription Workflow (Summary)

  1. Curator adds sources and resources to the Transcription Backlog.
  2. Resources from the backlog are sent to the AI Transcription Service for processing.
  3. AI-generated transcripts are published on the Registry and also added to the Review Queue.
  4. Reviewers claim and edit transcripts from the Review Queue and submit them for approval.
  5. Evaluators review and approve finalized transcripts.
  6. Approved transcripts are updated on the Registry as final versions, and Reviewers receive a payout for their work.

Transcription Workflow (Detailed)

  1. Curation: Curators propose new content by submitting PRs directly in the bitcointranscripts repository. This forms the Transcription Backlog, a collection of all the content that needs transcription, including specific sources like podcasts or individual resources like conference talks and YouTube videos.
  2. Generation & Queueing: The tstbtc transcription server processes sources from the Transcription Backlog and submits the AI-generated output back to the bitcointranscripts repository. This triggers a webhook in the transcription-review-backend, queuing the transcript for review and notifying via the Discord Server for its availability. Simultaneously, the transcription metadata are archived in the bitcointranscripts-metadata repository.
  3. Review: Reviewers access the transcription-review-front-end, an enhanced interface designed to streamline the transcript editing process. This interface is integrated with GitHub, allowing each edit to be tracked as changes in a unique branch of the reviewer's fork of the repository. The platform facilitates easy viewing, claiming, and modification of transcripts, with the backend coordinating the review workflow.
  4. Evaluation: After editing, reviewers submit their finalized transcripts, creating PRs to the bitcointranscripts repository. These PRs are then evaluated by an Evaluator. The Evaluator may request further revisions or approve the submission. Approved transcripts are merged into the main repository, marking the end of the review process.
  5. Payout: Following the approval, the transcription-review-backend processes the payouts for reviewers.
  6. Publication: Approved transcripts are merged into the main bitcointranscripts repository, automatically updating btctranscripts.com.

Components

  • Curator: Suggests source material for transcription.
  • bitcointranscripts: The central repository storing all transcripts and serving as the source of truth for the Transcription Backlog.
  • tstbtc: A transcription server that processes source material from the Transcription Backlog to produce AI-generated transcripts.
  • bitcointranscripts-metadata: An archive repository for raw model outputs and basic metadata generated by tstbtc.
  • transcription-review-backend: Coordinates the review workflow, managing transcript queues, notifications, and reviewer payouts.
  • transcription-review-front-end: A user interface for reviewers to manage their review tasks, including transcript editing and submission.
  • Discord Server: Community engagement including alerts for users about new transcripts available for review.
  • GitHub: Central platform for authentication, storage, and tracking submissions as part of the review and evaluation workflow.
  • Reviewer: Community contributors who refine the AI-generated transcripts, ensuring their quality and accuracy.
  • Evaluator: Responsible for the final review stage, assessing and approving submissions for merging.
  • Registry: The public-facing website for Bitcoin Transcripts, updated automatically with newly merged content.

CLICK for Sequence Diagram of the Transcription Pipeline
sequenceDiagram
    actor Curator
    participant tstbtc
    actor Evaluator
    participant bitcointranscripts
    participant Site as btctranscripts.com
    participant backend as transcription-review-backend
    participant discord as Discord Server
    participant GitHub
    participant frontend as transcription-review-front-end
    actor Reviewer
    
    Note over Curator,tstbtc: Transcript Creation Process
    opt transcription using tstbtc
        Curator->>+tstbtc: Submits source for transcription
        tstbtc-->>-Curator: Returns AI-generated transcript
    end
    Note over Curator,discord: Queueing Process
    Curator->>bitcointranscripts: Commits AI-generated transcript
    bitcointranscripts->>+backend: Triggers webhook for new content push
    backend->>backend: Adds new transcript to the Queue
    alt success
        backend-->>discord: Alerts users of new reviewable transcript
    else error
        backend-->>-discord: Alerts developers for error
    end

    Note right of bitcointranscripts: Repository Update
    bitcointranscripts->>Site: Triggers website rebuildte with new content

    Site->>Site: Rebuild with updated content

    Note over bitcointranscripts,Reviewer: Review Process
    Reviewer->>+frontend: Accesses review platform
    frontend->>+GitHub: Requests GitHub login
    GitHub-->>-frontend: Confirms authentication
    frontend-->>-Reviewer: Grants access to platform

    Reviewer->>+frontend: Requests transcripts list
    frontend->>+backend: Queries for available transcripts
    backend-->>-frontend: Returns available transcripts
    frontend->>-Reviewer: Displays transcripts

    Reviewer->>+frontend: Claims a transcript
    alt First-Time Claim
        frontend->>+GitHub: Fork bitcointranscripts repo
        GitHub-->>-frontend: Confirms repository fork
    end
    frontend->>+GitHub: Creates working branch
    GitHub-->>-frontend: Confirms branch creation
    frontend->>+backend: Registers transcript claim
    backend-->>-frontend: Return assigned review ID
    frontend->>-Reviewer: Shows claimed transcript for editing

    loop Edit Transcript
        Reviewer->>+frontend: Save edits
        frontend->>+GitHub: Commit edits to branch
        GitHub-->>-frontend: Confirms commit
        frontend-->>-Reviewer: Save confirmation
    end
    
    Reviewer->>+frontend: Submits Review
    frontend->>+GitHub: Initiates Pull Request
    GitHub->>+bitcointranscripts: Opens PR to bitcointranscripts repo
    GitHub-->>-frontend: PR creation confirmation
    frontend->>+backend: Notifies backend of submission
    backend-->>-frontend: Acknowledges submission
    frontend-->>-Reviewer: Confirms submission

    Note over Evaluator,Reviewer: Evaluation Process
    loop PR Evaluation
        Evaluator->>+bitcointranscripts: Reviews PR
        alt Comments
            bitcointranscripts-->>Reviewer: Requests change
            loop Edit Transcript
                Reviewer->>+frontend: Save edits
                frontend->>+GitHub: Commit edits to branch
                GitHub-->>-frontend: Confirms commit
                frontend-->>-Reviewer: Save confirmation
            end
        else Accept
            bitcointranscripts-->>Reviewer: Approves PR
        end
    end

    Evaluator->>bitcointranscripts: Merges PR
    bitcointranscripts->>backend: Triggers webhook for PR merge
    backend->>backend: Calculate payout and credit to Reviewer's account

    Note right of bitcointranscripts: Repository Update
    bitcointranscripts->>Site: Triggers website rebuild
    Site->>Site: Rebuild with updated content

Loading

Popular repositories Loading

  1. bitcointranscripts bitcointranscripts Public

    A treasure trove of transcripts associated with Bitcoin and Lightning Network

    Python 94 195

  2. transcription-review-backend transcription-review-backend Public

    TypeScript 7 11

  3. tstbtc tstbtc Public

    This cli app transcribe audio and videos for submission to the bitcointranscripts repo

    Python 6 8

  4. bitcointranscripts.github.io bitcointranscripts.github.io Public

    A static Hugo site for hosting btctranscripts.com

    SCSS 5 13

  5. registry registry Public

    TypeScript 4 3

  6. transcription-review-front-end transcription-review-front-end Public

    TypeScript 3 7

Repositories

Showing 8 of 8 repositories