A cloud-hosted persistent memory server for MCP agents — store text, docs, and audio, then recall it in plain English or with audio-specific filters.
remember — ingest raw text, a document/PDF URL, or an audio URL (transcribed via Whisper); auto-returns a summary, action items, and key topics.
recall — ask natural-language questions of stored memory and get an AI-synthesized answer with source citations, including across past sessions.
note — save a quick tagged observation, preference, or decision as a searchable vector (faster than remember).
namespace — report which namespace the server reads/writes (default nl), for isolating work/personal projects.
audio_ingest — process calls, voice notes, podcasts, meetings, music, or environmental audio into semantic memory (requires BLUECOLUMN_API_URL).
audio_recall — search audio memories with type filters (preference, promise, fact, action_item) and top_k.
call_prepare / call_complete — pull relationship context before a voice call (unresolved items, suggested opening, delivery preferences) and store what changed after it (follow-ups, sentiment).
sound_analyze — detect and index non-speech audio events.
music_analyze — extract tempo, structure, instrumentation, energy curve, and mood from music.
Note: the README also advertises music_remember, music_recall, streaming_audio_ingest, and streaming_audio_recall, but these are not present in the server schema.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@bluecolumn-mcp What did we decide about pricing?"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Give Claude Desktop, Cursor, and any MCP-compatible agent persistent memory across sessions.
2-minute setup. Zero-config. Audio-native. Works with every MCP client.
MCP server for BlueColumn — persistent semantic memory for AI agents.
Give any MCP-compatible agent (Claude Desktop, Cursor, Windsurf, LangChain, AutoGen, CrewAI) the ability to remember, recall, and store observations across sessions.
Tool | Description |
| Ingest text, audio, or documents into persistent memory |
| Query memory with natural language, get AI-synthesized answer + sources |
| Store lightweight agent observations as searchable vectors |
| Show the configured namespace |
Tool | Description |
| Store a musical recording, practice session, or lesson with rich musical context — instrument, key, tempo, technique tags, chord progression, notes. Audio is transcribed; the structured context is embedded alongside so a coach/teacher agent recalls it with full musical understanding |
| Search stored musical content with musical filters — instrument, technique, or style. E.g. "show me every take with barre chord issues" or "what did we practice in E minor last month" |
Tool | Description |
| Ingest calls, voice notes, podcasts, meetings, or music and extract semantic memory |
| Search audio memories and get audio-backed citations |
| Get memory context before a voice call starts |
| Store what changed after a call ends |
| Detect and index non-speech audio events |
| Extract tempo, structure, instrumentation, mood from music |
Tool | Description |
| Ingest edge-device audio chunks (car, doorbell, wearable) into per-device streaming memory |
| Recall over a device's streamed audio history |
Related MCP server: hypermnesia
Store a practice session:
music_remember {
"audio_url": "https://example.com/take3.mp3",
"title": "Week 3 practice — barre chords, Sep 7",
"instrument": "guitar",
"musical_key": "E minor",
"tempo_bpm": 92,
"technique_tags": ["barre-chords", "hammer-on"],
"style_tags": ["blues"],
"chord_progression": ["Em", "Am", "B7"],
"notes": "F-shape barre still buzzy on the B string; Fmaj7 subst worked well"
}Then recall it musically:
music_recall { "q": "which takes had barre chord issues?", "instrument": "guitar" }Sign up free at bluecolumn.ai — 60 min audio + 100 queries/month, no credit card required.
npm install -g bluecolumn-mcpAdd to ~/Library/Application Support/Claude/claude_desktop_config.json:
{
"mcpServers": {
"bluecolumn": {
"command": "bluecolumn-mcp",
"env": {
"BLUECOLUMN_API_KEY": "bc_live_YOUR_KEY"
}
}
}
}Done. Claude now remembers everything you tell it to.
Feature | BlueColumn MCP | Other MCP Memory Servers |
Audio ingestion | ✅ Native (transcribe + store) | ❌ Text only |
Plain English recall | ✅ "What did we decide about pricing?" | ❌ Keyword/vector only |
Setup time | ~2 minutes | ~10-30 minutes |
Cloud managed | ✅ No infra to run | ⚠️ Most require local DB |
Namespace isolation | ✅ Built-in per key | ⚠️ Manual |
Remember: We're building a React component library called Aether with Tailwind CSS v4.What project was I working on? What was my tech stack?→ Claude recalls: "You're building Aether, a React component library with Tailwind CSS v4"
Record a voice memo → BlueColumn transcribes it → Claude remembers it.
{
"mcpServers": {
"bluecolumn": {
"command": "bluecolumn-mcp",
"env": {
"BLUECOLUMN_API_KEY": "bc_live_YOUR_KEY",
"BLUECOLUMN_NAMESPACE": "my_project"
}
}
}
}Use different namespaces for work, personal, and side projects. Each stays isolated.
NEW (Audio Intelligence): Ingest audio (calls, voice notes, podcasts, music) and extract semantic memory using the audio-intelligence layer. Requires BLUECOLUMN_API_URL to point at the BlueColumn API.
Inputs: audio_url (required), customer_id (required), source_type (required), metadata.
NEW (Audio Intelligence): Search audio memories and get audio-backed citations.
Inputs: customer_id (required), query, top_k, filter_type.
NEW (Audio Intelligence): Store what changed after a call ends.
Inputs: call_id (required), customer_id (required), new_memories, follow_up_required, +1 more.
NEW (Audio Intelligence): Get memory context before a voice call starts.
Inputs: customer_id (required), include.
NEW (Audio Intelligence): Extract tempo, structure, instrumentation, mood from music.
Inputs: audio_id (required), extract.
Get the BlueColumn namespace this MCP server is configured to write/read (default: nl).
Inputs: no named parameters.
Store a lightweight agent observation as a searchable vector. Use when the agent wants to save a quick preference, decision, or observation without needing full document processing. Faster than remember for short notes.
Inputs: text (required), tags.
Query BlueColumn memory using natural language. Returns an AI-synthesized answer with source citations. Use when the agent needs to retrieve past information, answer questions about stored content, or get context from previous sessions.
Inputs: q (required).
Store text, a document URL, or audio URL into BlueColumn persistent memory. Returns a summary, action items, and key topics automatically extracted by AI. Use when the user or agent wants to save information for future recall.
Inputs: text, title, file_url, audio_url.
NEW (Audio Intelligence): Detect and index non-speech audio events.
Inputs: audio_id (required), search_for, detect_events.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 10 tools
Tools like note, remember, call_complete, and audio_ingest all store information, with subtle distinctions that could lead to misselection. recall and audio_recall also overlap, though audio_recall is more specific.
Names mix conventions: some are verbs (remember, recall), some are noun_verb (audio_ingest, sound_analyze), and some are verb-like but structured as object+verb (call_prepare, call_complete). No consistent pattern.
10 tools is a reasonable count for the server's apparent scope. While a few tools feel peripheral (sound_analyze, music_analyze), the overall number is not excessive.
The server covers storage and retrieval well, including audio-specific features. However, there is no delete or update operation for memories, which is a notable gap in lifecycle management.
Persistent memory for AI agents across Claude, ChatGPT and any MCP client.
Persistent memory for AI agents to retain, retrieve, and recall conversation context through MCP.
Persistent memory, knowledge graph, task management and more for AI agents via MCP.
Persistent personal memory for AI assistants — save, search, and recall across every MCP client.
