Voice

Audio Memory

Name: Audio Memory
Rating: 3.5 (1 reviews)
Author: bluecolumnconsulting-lgtm

By bluecolumnconsulting-lgtm 👁 107 views ▲ 0 votes

BlueColumn Audio Memory — OpenClaw plugin: audio memory, call intelligence, and Music Memory (store/recall practice recordings with instrument, key, tempo, technique, chord context)

GitHub

Configuration Example

{
  "apiKey": "bc_live_...",
  "baseUrl": "https://xkjkwqbfvkswwdmbtndo.supabase.co/functions/v1",
  "audioApiUrl": "http://localhost:8000",
  "namespace": "nl"
}

README

# BlueColumn Audio Memory

Persistent audio memory and call intelligence for OpenClaw.

BlueColumn Audio Memory turns calls, voice notes, meetings, podcasts, and music into searchable memory and structured audio intelligence.

It is built for OpenClaw agents that need to remember what happened in voice conversations, prepare before the next call, and retrieve audio-backed context later.

## What it does

- Ingest audio from calls, voice notes, meetings, podcasts, and music
- Search audio memories with citations
- Prepare an agent before a call with relevant relationship context
- Store what changed after a call ends
- Detect non-speech sound events
- Analyze music structure and metadata
- Track sentiment trends across calls
- Generate CRM-style summaries from audio memory

## Core use cases

- Sales agents that need to remember promises, objections, and follow-ups
- Support agents that need customer continuity across calls
- Founder assistants that track conversations and decisions
- Recruiter workflows that need candidate call memory
- Agencies managing many client conversations
- Voice-first AI systems that need durable memory

## Included tools

### `bluecolumn_music_remember`
Store a musical recording, practice session, or lesson with rich musical context — instrument, key, tempo, technique tags, chord progression, and free-form notes. Audio is transcribed; the structured context is embedded alongside so recall understands the music, not just the words.

### `bluecolumn_music_recall`
Search stored musical content with musical filters — by instrument, technique, or style. Ask things like "show me every take with barre chord issues" or "what did we practice in E minor last month."

### `bluecolumn_audio_ingest`
Ingest audio and extract semantic memory.

### `bluecolumn_audio_recall`
Search audio memories and retrieve audio-backed context.

### `bluecolumn_call_prepare`
Get memory context before a voice call starts.

### `bluecolumn_call_complete`
Store what changed after a call ends.

### `bluecolumn_sound_analyze`
Detect and index non-speech audio events.

### `bluecolumn_music_analyze`
Extract tempo, structure, instrumentation, energy, vocal characteristics, and motifs from music.

### `bluecolumn_sentiment_trend`
Summarize sentiment trend over calls for a customer.

### `bluecolumn_crm_summary`
Generate CRM-style summaries from call and audio memory.

## Why this plugin is different

This is not just transcription and it is not just TTS.

BlueColumn Audio Memory is designed to give OpenClaw agents an audio memory layer:

- searchable call memory
- call preparation
- post-call memory updates
- music ingestion with instrument, key, tempo, technique, and chord context
- music memory recall filtered by instrument, technique, or style
- sound-event intelligence
- music intelligence
- sentiment trends
- CRM-style summaries

## Configuration

Example plugin config:

```json
{
  "apiKey": "bc_live_...",
  "baseUrl": "https://xkjkwqbfvkswwdmbtndo.supabase.co/functions/v1",
  "audioApiUrl": "http://localhost:8000",
  "namespace": "nl"
}
```

Environment variables supported:

- `BLUECOLUMN_API_KEY`
- `BLUECOLUMN_API_URL`
- `BLUECOLUMN_NAMESPACE`

## Who should install this

Install this when your OpenClaw agent needs to:

- remember calls across sessions
- search past voice conversations
- capture action items and follow-ups from audio
- prepare for the next call automatically
- understand non-speech audio events
- analyze music or structured audio

## Bookmark-worthy reason

If you expect to build sales, support, voice, or customer-memory agents later, this is the plugin to save now.

Your agent should remember voice conversations the same way it remembers text.

## Status

Early scaffold. Some advanced features may start as partial implementations before full release hardening.

voice