What it is
A web and Android-first workspace for capturing calls and turning them into durable, searchable operational memory.
02 / Product / applied AI
Membrain
A memory workspace that turns recorded conversations into transcripts, decisions, tasks and a connected knowledge graph.
Launch live demo ↗Active build
Focus
- Product
- Audio intelligence
- Knowledge graph
Project conversationaudio · captured today
TRANSCRIPT
We should connect the deployment notes to the project…
I’ll turn that into a task and keep the context attached.
EXTRACTED
DecisionShip through the internal stack
Next actionDocument the workflow
audio→transcript→memory→action
PROJECT NOTES
A technical case study of the product, its working model and its current state.
Problem
Important decisions and follow-ups disappear inside recordings. The useful context is hard to retrieve and the resulting work is easy to lose.
Approach
Treat every recording as the start of a pipeline: ingest audio, transcribe it, analyze the conversation, materialize tasks and update a tenant-scoped knowledge graph.
What I built
Web and mobile capture, provider-selectable transcription, structured analysis, speaker assignment, task materialization, people directories and local-focus graph exploration.
HOW IT WORKS
From input to useful output.
Capture
WebAndroidAudio
→Process
TranscribeAnalyze
→Materialize
TasksPeopleGraph
- 01Record / upload
- 02Transcribe
- 03Analyze
- 04Connect
- 05Act
SYSTEM / STACK
The pieces behind the interface.
Current stack
- React
- Flutter
- Firebase
- Cloud Functions
- OpenAI
- AssemblyAI
CURRENT STATE