Skip to main content
Glama
Sign In
deenesjakoruzh
Samuraizer

Samuraizer

by UladzKha
OverviewSchemaSourceRelated ServersDiscussionsAdmin
TypeScript
Local

Samuraizer

Turn meeting recordings into transcripts, summaries, action items, and decisions — entirely on your machine. No cloud, no subscriptions, no data leaving your network.

Reference implementation. Samuraizer is the reference implementation of the memnex specification — an open standard for portable meeting outputs (transcripts, summaries, action items, decisions). All outputs conform to memnex v0.2, including a full provenance chain.

Why Samuraizer

  • Fully local. Your recordings never leave your machine.

  • CLI-first. Scriptable, automatable, integrates with cron, Git hooks, Obsidian workflows.

  • Resumable. Crashed mid-pipeline? Re-run picks up where it left off.

  • Fast. LLM stages (summary, action items, decisions) run in parallel, cutting total time by ~2.3×.

  • Model-agnostic. Works with any Ollama-compatible LLM — pick what fits your hardware.

  • Free. No subscriptions, no per-minute pricing.

Related MCP server: m4x

Portable output with memnex

Samuraizer writes meeting results in the portable memnex v0.2 format. This lets another meeting app, local AI agent, or MCP server consume the same transcript, summary, action items, decisions, and provenance without learning a Samuraizer-specific JSON shape.

To verify a memnex document in under a minute, follow the memnex quickstart. It installs the validator and checks a real v0.2 example end to end.

📦 Packages

This is a monorepo. Samuraizer is published as two npm packages, each with its own README and changelog:

Package

What it does

Docs

@samuraizer/cli

The meeting-processing CLI — install, configure, and run the local pipeline. Start here.

packages/cli/README.md

@samuraizer/mcp-server

Companion Model Context Protocol server. Lets AI agents (Claude Desktop, Claude Code, MCP Inspector) query your processed meetings and trigger the pipeline.

packages/mcp-server/README.md

🏛 Project status

  • memnex specification: v0.2 — see the memnex repo for the schema, conformance suite, and governance documents.

  • Schemastore registered. memnex is in the universal JSON Schema catalog (PR #5676). *.memnex.json and meeting-output.json files get IDE autocomplete and validation by default in VS Code, JetBrains IDEs, Neovim, Sublime, and other editors with language-server support.

  • Roadmap: local diarization, cryptographic signing UX, accessibility (WebVTT subtitles), and the path to memnex v1.0 are funded milestones.

🤝 Contributing

Samuraizer follows a BDFL bootstrap model — see the memnex GOVERNANCE.md for the broader project's governance model, which Samuraizer mirrors.

Bug reports and feature discussions are welcome in GitHub Issues.

📝 Changelogs

📄 License

MIT — see LICENSE.

Available Tools

9 tools
extract_action_itemsExtract Action ItemsA

Extract action items from a meeting transcript. Returns a JSON list of tasks with owner and due date.

Inputs: transcriptText (required), model.

extract_decisionsExtract DecisionsB

Extract confirmed decisions from a meeting transcript. Returns a JSON list of decisions.

Inputs: transcriptText (required), model.

get_meetingB

Retrieve the full processed output for a specific meeting by id.

Inputs: id (required).

list_meetingsA

List all processed meetings, sorted newest first. Optionally limit results.

Inputs: limit.

normalize_audioNormalize AudioB

Normalize an audio file to 16kHz mono PCM WAV format required by Whisper.

Inputs: inputPath (required), outputPath (required).

process_recordingProcess RecordingB

Run the full Samuraizer pipeline on an audio file. Returns summary, action items, decisions, and output file paths.

Inputs: filePath (required), model.

search_meetingsA

Search meetings by summary text, name, action items, and decisions. Transcripts are not searched — use get_meeting for those. Returns ranked results with snippets.

Inputs: query (required), limit.

summarize_transcriptSummarize TranscriptC

Generate a concise meeting summary from transcript text.

Inputs: transcriptText (required), model.

transcribe_audioTranscribe AudioB

Transcribe an audio file using whisper.cpp. Returns the transcript text.

Inputs: filePath (required).

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Deploy Server
    One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

    TDQS

    A3.8/5.0

    Scored across 9 tools

    Disambiguation5/5

    Each tool has a clearly distinct purpose, from normalization and transcription to summarization, extraction, and meeting management. The separate extraction tools for action items and decisions avoid overlap, and process_recording serves as a composite operation without ambiguity.

    Naming Consistency5/5

    All tool names follow a consistent verb_noun pattern in snake_case (e.g., normalize_audio, transcribe_audio, list_meetings), making it predictable and easy to understand the function of each tool.

    Tool Count5/5

    With 9 tools, the count is well-suited to the domain of audio processing and meeting summaries. It covers the essential steps of the pipeline and meeting retrieval without being excessively large or minimal.

    Completeness5/5

    The tool set covers the full workflow from raw audio normalization to transcription, summarization, extraction of action items and decisions, and includes a composite processing tool. Retrieval, search, and listing functions complete the necessary operations for managing processed meetings.

    Maintenance

    ActivityStale
    ResponsivenessNo recent community issues

    Resources

    Related MCP Connectors

    Related MCP Servers

    • STT2TTS MCP

      pygodzilla
      A
      license
      Not graded
      quality
      D
      maintenance
      Local-first speech-to-text and text-to-speech MCP server. Hot-swappable engines via config.yaml — no code changes, no API keys required.
      2
      MIT

    • m4x

      closeofbusiness
      A
      license
      A
      quality
      D
      maintenance
      Local, private audio transcription MCP server enabling AI agents to transcribe audio files entirely on-device without uploading data.
      3
      MIT

    • meetey

      jej2k5
      A
      license
      Not graded
      quality
      C
      maintenance
      Local-first meeting capture and transcription for Claude Code. Records audio from meeting apps, transcribes locally with whisper.cpp, and produces structured notes via Claude.
      1
      Apache 2.0

    • asr-mcp

      fengredrum
      A
      license
      A
      quality
      D
      maintenance
      MCP server for local speech-to-text using Whisper Large V3 (MLX), enabling audio transcription with text/timestamps/SRT output and LLM-based correction, all running offline on Apple Silicon.
      2
      MIT