Feat/aigateway transcription no UI#5
Open
JJassonn69 wants to merge 34 commits intonextfrom
Open
Conversation
… and add App Dev Server
- Add .mp4 and .mkv to .gitignore - Update vscode launch configs to use node debugger with bash for better process control - Add compound launch config for running streamplace + app dev server together - Add subtitle toggle to viewer context menu - Implement subtitle support in HLS player with proper track management - Change default player protocol from WebRTC to HLS - Add aigateway package for AI transcription gateway client with session
- Add subtitle offset controls to viewer context menu with +1s/-1s/reset options - Display current subtitle offset in menu title and adjustment UI - Pass subtitle offset as query parameter to HLS source URL - Add subtitleOffsetMS state to player store with getter/setter - Implement structured transcript segments with timing in aigateway - Add TranscriptSegment and WordTimestamp types for precise subtitle timing - Refactor TranscriptStore to use segments
This reverts commit cc2b68b.
Refactor AI gateway client to try multiple endpoint paths for better compatibility with different gateway versions. Remove RewriteURLsTo config option in favor of automatic URL normalization based on the configured base URL. Add StopURL field to Session response. Update TranscriptEvent to use RFC3339 timestamp format instead of milliseconds. Remove unused error tracking from RTMPPublisher and legacy transcript segment conversion logic.
Remove SegmentAndSignElemH264Parse function which was a duplicate of SegmentAndSignElem with identical functionality. Update WebRTCIngest to use the standard SegmentAndSignElem method instead.
Add debug segment dump when MP4 validation fails after signing to help diagnose segment validation issues.
Remove transient mux error retry logic and excessive timestamp zeroing from ConvergeSegment. Add debug segment dump at start of convergence and per-attempt debug file output when SegmentDebugDir is configured. Simplify MP4 metadata zeroing to only handle bitrate fields instead of all timestamps.
Decrease pipeline start timeout in MP4ToMPEGTS from 60 seconds to 10 seconds to fail faster when conversion stalls. Add missing newline at end of segment_converge.go.
Remove AIGatewayPathPrefix config option and related PathPrefix field from aigateway.Config. Simplify endpoint discovery to only try base paths without prefix variants, removing the /gateway/ path candidates that were previously used for proxy compatibility.
Remove the Text field from TranscriptEvent struct in favor of structured Segments data. Simplify SSE payload parsing to expect single-event format wrapped in array instead of handling multiple legacy formats. Update debug logging to remove text field reference.
Remove automatic deduplication of transcript segments by eliminating the seen map tracking and stableSegmentID generation. Remove segment replacement logic that attempted to update segments with similar timestamps. Remove sorting of segments in GetSegments and GenerateVTTForSegment, relying on insertion order instead. Simplify addSeg to only enforce MaxTranscriptEvents limit without deduplication checks.
…ayer components - Eliminated subtitle display logic and offset management from the context menu and video player. - Updated player state and store to remove subtitle-related properties and methods. - Adjusted media handling to focus on WebRTC protocol, removing RTMP dependencies. - Cleaned up AI gateway configurations by removing unused RTMP settings.
eliteprox
reviewed
Jan 8, 2026
eliteprox
left a comment
eliteprox
left a comment
There was a problem hiding this comment.
We should also explore moving aigateway into a go module in separate muxionlabs org repo for reusability
| base := strings.TrimRight(cfg.BaseURL, "/") | ||
| startCandidates := []string{ | ||
| base + "/process/stream/start", | ||
| base + "/ai/stream/start", |
There was a problem hiding this comment.
can be removed when using latest muxionlabs/go-livepeer:transcoding-with-ai
There was a problem hiding this comment.
We can try moving this into segment_conv.go to see if it will cover WHIP ingest also. See https://github.com/streamplace/streamplace/pull/810/changes#diff-8d7219fc0d142be71d6d322c932bbd3cc7c63c1a11a7d8d4cc942138f753d8bd
- Removed the aigateway package and its related files, streamlining the codebase. - Updated go.mod to reflect changes in dependencies, including the addition of muxionlabs/ai-go-sdk. - Adjusted launch configuration for local development to use the correct AI gateway URL. - Enhanced media handling by integrating the new ai-go-sdk for AI gateway interactions.
- Updated go.mod to use version 0.1.1 of muxionlabs/ai-go-sdk. - Enhanced LivestreamState to include transcripts, allowing for real-time AI-generated text segments. - Implemented handling of transcript segments in websocket consumer, ensuring they are stored and published correctly. - Added new lexicon definitions for transcript segments and words to support structured data handling. - Refactored media ingestion to initiate AI sessions and process transcript events, improving the overall media management workflow.
- Replaced AI gateway stream configuration with a new structure for better clarity. - Increased buffer sizes for audio and video channels to improve performance. - Adjusted pipeline configuration for video processing to streamline data flow. - Removed redundant context handling for transcript streaming, ensuring cleaner code.
- Introduced AISessionResources structure to encapsulate video and audio channels along with cleanup functionality. - Simplified transcript event handling by centralizing the publishing logic in PublishTranscriptToBus. - Updated MKV and RTMP ingest methods to utilize the new AISessionResources for improved clarity and maintainability. - Removed redundant code related to transcript event processing, enhancing overall code cleanliness.
refactor: transcription sdk
![cursor[bot]](https://avatars.githubusercontent.com/in/1210556?s=60&v=4){kind=small}
There was a problem hiding this comment.
This PR is being reviewed by Cursor Bugbot
Details
You are on the Bugbot Free tier. On this plan, Bugbot will review limited PRs each billing cycle.
To receive Bugbot reviews on all of your PRs, visit the Cursor dashboard to activate Pro and start your 14-day free trial.
…m/muxionlabs/streamplace into feat/aigateway-transcription-noUI
…ogic, remove excessive documentation and follow the code pattern in the repository
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No UI changes, only the essential for adding aigateway.
Notion Instructions instructions for testing locally.
Note
Adds backend AI transcription via
github.com/muxionlabs/ai-go-sdkwith channel-driven GStreamer integration and event publishing.AISessionResources,SetupAISession, andNewAISinkCallbackinpkg/media/*, plusaigateway_loggerandTranscriptStoreonMediaManagerplace.stream.ai#dataOutput--ai-gateway-base-urland--ai-gateway-pipelineflags to wire up the gatewaywebsocket-consumer.tsxlogs AI data output to console (no visual changes).gitignorefor video files, andutil/live_mkv_with_new_did.shhelperai-go-sdk; bumppion/webrtcand related libsWritten by Cursor Bugbot for commit d120b76. This will update automatically on new commits. Configure here.