OpenPlod
Local-first Plaud Note Pro audio vault. Direct Bluetooth imports, Mistral/Voxtral transcription, Markdown documents, AI chat, REST API and MCP. Tauri macOS app; experimental Android.
Install / Use
claude mcp add RemiPelloux -- npx -y github:RemiPelloux/OpenPlodIf the server publishes to npm under a different name, use that package instead — check the repo README.
MCP Server
Model Context Protocol server
Quality Score
Category
Content & MediaSupported Platforms
Our assessment of OpenPlod
OpenPlod scores 78/100 on our quality scale, 1035th of 1,153 Content & Media skills we index.
Its MCP Server is 42 KB long, well organised into 31 sections with 11 code examples: long enough that it reads more like full documentation than a focused instruction file, which agents can find harder to follow.
It has 46 GitHub stars, so there is little community track record yet; judge it on its content.
Maintenance, license and trust
- The repository was last updated yesterday, so OpenPlod is actively maintained.
- Our last check on 2026-09-25 found the source still online.
- It is released under the MIT license, a permissive license that allows use, modification and commercial use with attribution.
- Its trust signals score 97/100, with no cautions. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.
Safety scan
No issues foundOur scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful.
AI review by kimi-k2.7-code on 2026-09-24. Automated pattern scan on 2026-09-24. It catches known dangerous patterns, not every risk — read a skill before letting an agent act on it.
OpenPlod compared with similar skills
All 4 of these similar skills score higher than OpenPlod; compare them before choosing.
| Skill | Score | Stars | Updated | Format |
|---|---|---|---|---|
| OpenPlod (this skill)by RemiPelloux | 78 | 46 | 1d ago | MCP Server |
| Agent-Reachby Panniantong | 100 | 112.3k | 9d ago | CLAUDE.md |
| rufloby ruvnet | 100 | 81.9k | today | CLAUDE.md |
| headroomby headroomlabs-ai | 100 | 79.8k | today | CLAUDE.md |
| CowAgentby zhayujie | 100 | 48.0k | today | CLAUDE.md |
Frequently asked questions
- How do I install OpenPlod?
- Run
claude mcp add RemiPelloux -- npx -y github:RemiPelloux/OpenPlod. The install tabs above show the steps for each supported agent. - Which AI agents does OpenPlod work with?
- It is written for Claude Code and Claude Desktop, as a MCP Server file. Other agents that read the same format can often use it too.
- Is OpenPlod safe to use?
- Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful. It is MIT-licensed and scores 97/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
- Is OpenPlod still maintained?
- The repository was last updated yesterday, so OpenPlod is actively maintained.
Skill content
View source on GitHubOpenPlod brings recording, playback, transcription, and export into one desktop-first workspace. The desktop app owns the durable audio vault on macOS, Linux, and Windows; the Android companion pairs over your private network and can extract directly from an enrolled Note Pro. Keep recordings on your Plaud, keep a copy on your computer, and work with transcripts as documents.
[!IMPORTANT] 0.6.0 is an experimental prerelease, not a plug-and-play Plaud replacement. Direct Bluetooth extraction has been verified on one authorized Plaud Note Pro. Other devices and firmware are not established. The recorder still refuses audio until this computer holds a Plaud-minted identity, so authorization is required on every platform. You can use the recording vault, microphone, file imports, transcripts, and documents without a Plaud identity.

Actual application and local vault, with the Export tab selected. Private recording titles and tags are masked; no mock records or fabricated processing results are shown.
New in 0.6.0
Transcript Studio — a new Studio mode on the recording page, with five tabs:
- Edit — speaker rename/merge, segment split/merge, and find/replace with match-case, whole-word and per-speaker scoping. Preview before applying; undo restores the previous version.
- Versions — a word-level diff that labels each change as your correction, a regeneration, your own revision, or model output replacing your edit. Promotion never discards the replaced version.
- AI cleanup — punctuation, paragraphs, optional filler removal and headings, shown as a diff and saved only when you accept. Translation is stored as its own lineage and never replaces the original.
- Structure — chapters, decisions, action items and open questions, with SRT/VTT subtitles and Markdown, CSV, JSON and print-ready exports.
- Actions — save an instruction once and re-run it on any transcript.
Also: playback following with word-level highlighting where a provider supplies word timings, seven document templates (notes, meeting minutes, interview, lecture, brief, PRD, follow-up email), and batch transcription with per-item status, cancellation and retry-failed-only.
What it refuses to do is the point. Models are never allowed to emit a timestamp — they cite a segment index and the time is read from your own data, so an invented citation produces no link rather than a convincing-looking time. An unstated owner or due date stays unknown instead of being guessed. Subtitles are refused, with the reason, when a provider returned no real timing. Nothing is written on a first click, and a batch tells you what it will run before it runs it.
Not verified: the 0.6.0 exit gate — real-audio acceptance across multiple speakers, French and English, silence, noise and a long recording — has not been run, and no AI path has been exercised against a live provider. See the 0.6.0 changelog for the full list.
Previously in 0.5.1
0.5.1 was a platform patch on 0.5.0 that made Bluetooth cross-platform.
- One Bluetooth bridge on every desktop.
plaud-bridge(Rust +btleplug) is now the default on macOS, Linux, and Windows, binding CoreBluetooth, BlueZ, and WinRT under one protocol. The macOS Swift helpers remain only as an opt-in fallback (OPENPLOD_BLE_BACKEND=swift) and are never selected automatically. - Host autodetection. The Plaud page has a "This computer" panel that checks the platform, bridge, adapter power,
ffmpeg/ffprobe, and recorder authorization separately, and names the fix for the platform it is running on. Previously any one of these failing showed the same "Bluetooth scan failed". bun run doctor. The same report from a terminal, with--jsonfor bug reports. It exits non-zero when something blocks direct transfer, so it works as a setup gate.plaud-bridge doctor. A non-scanning host check that reports the backend, adapter, and power state. It succeeds even with no adapter, so the vault renders a diagnosis instead of an error.- Scan reliability. The scan probe now gets the same connect budget as the transfer path. A cold BlueZ cache made the first scan after boot report a healthy recorder as unreachable.
- Linux and Windows packaging.
.deb/.rpmbundles declare their BlueZ, D-Bus, and ffmpeg dependencies, and CI builds and lints the bridge on Linux, macOS, and Windows.
Verified on Linux with real hardware: an authorized Plaud Note Pro (protocol 20) was discovered, connected, and its GATT command service opened on Arch Linux via BlueZ. No new device models or firmware are claimed, and no end-to-end Linux audio download is claimed — that still needs a provisioned identity.
Previously in 0.5.0
- Separate transcription, summary, document, and chat provider settings. Existing Mistral keys are retained.
- OpenAI and AssemblyAI speech adapters; OpenAI, Anthropic, and local Ollama text adapters.
- Per-recording transcription provider/model selection, language detection, supported vocabulary hints, and explicit fallback.
- Local-only processing, persisted jobs, cancellation, known-remote-job resume, and generated transcript provenance.
- Shared reviewed Markdown generation and cited chat, with provider-specific consent labels.
Experimental, not production-accepted: new cloud providers have contract tests but no live acceptance with owner credentials; Ollama was not running on the test Mac. Spending controls, automatic long-audio chunking, complete per-model language catalogs, and production Android acceptance remain unfinished. Large OpenAI files are rejected before upload, not silently truncated. See 0.5 implementation status and the 0.5.0 prerelease.
Plaud compatibility has not expanded: direct extraction was verified only on the documented authorized Note Pro setup. Plaud Note, NotePin, other models, and untested firmware are unverified, not universally supported.
Previously in 0.4.1
- Transcript workspace: compact list and Markdown reader, export menus, editing links, and Mistral document creation.
- AI Chat: collapsible source/history controls, selected-source chips, prompt shortcuts, answer copying, and a cleaner composer with explicit cloud consent.
- Android connections: a dedicated page with a private pairing QR that stays hidden until requested, vault reachability checks, and confirmed unpairing.
- Navigation and reliability: small-screen transcript Back navigation, guarded asynchronous responses, and real-library regression checks.
- Road to 1.0: a phased feature roadmap with additional AI-provider targets and measurable release gates. Provider adapters were added subsequently in the experimental 0.5.0 release.
See the 0.4.1 changelog for validation and limitations.
Previously in 0.4.0
- A consistent workspace: redesigned Recordings, Documents, and Plaud Device pages; original SVG icons; compact toolbars; light/dark themes.
- AI with selected context: choose recordings, ask Mistral questions, follow source references, and export the conversation as Markdown.
- Better audio workflows: reliable loading, playback speed, bookmarks, segment corrections, and tags from a recording's right-click menu.
- A document editor: Markdown formatting, a heading outline, transcript provenance, revision history, folders, and export controls.
- Direct-device improvements: opt-in automatic import, interrupted Bluetooth transfer checkpoints, and an SDK-free Android adapter in the source tree.
- Verifiable Mac updates: a visible build date and an installer that archives duplicate bundles instead of leaving several launchable copies.
See the changelog for validation and remaining limitations.
Road to 1.0.0
The 1.0.0 roadmap expands OpenPlod into a multi-provider transcript and knowledge workspace while keeping original audio safe. The 0.5.0 prerelease delivers an initial implementation with acceptance gates still open; later milestones remain planned:
- 0.5: OpenAI and AssemblyAI speech-to-text, stronger Mistral/Deepgram/local Whisper controls, and separate cloud/local AI choices for transcript analysis.
- 0.6: Transcript Studio with reviewed AI cleanup, translation, version comparison, chapters, reusable document templates, and subtitle exports.
- 0.7: Linked Markdown knowledge, managed Obsidian export, richer source-linked chat, and expanded API/MCP automation.
- 0.8: Resumable Android uploads, offline synchronization, pairing management, and validated direct-device recovery.
- 0.9 -> 1.0: Fresh-owner Plaud onboarding, backup/restore, privacy hardening, performance validation, and signed releases.
Additional transcript-analysis targets include OpenAI, Anthropic, and local Ollama; speech-to-text and text analysis remain separate capabilities. New providers require explicit configuration, consent, and real-data validation. See the provider plan and release gates.
Download
| Platform | Package | Notes | | --- | --- | --- | | macOS Apple Silicon | OpenPlod 0.5.0 ZIP | Experimental. Ad-hoc signed, not notarized. Requires external tools for direct Plaud extraction. | | Android ARM64 | Build instructions | Current source uses a native, SDK-free direct Bluetooth adapter. Development build only; no new public APK release yet. | | Source | [Tagged source and release notes](https://github.com/RemiPello
Truncated for display — read the full file on GitHub.
Related Skills
Agent-Reach
112.3kGive your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees.
ruflo
81.9k🌊 The original agent harness. Deploy intelligent multi-player swarms, coordinate autonomous workflows, and build conversational AI systems. Features adaptive memory, self-learning intelligence, federation, vector RAG integration, and native Claude Code / Codex / Hermes and many more Integrated
headroom
79.8kCompress tool outputs, logs, files, and RAG chunks before they reach the LLM. 20% fewer tokens for coding agents, 60-95% fewer tokens for JSON, same answers. Library, proxy, MCP server.
CowAgent
48.0kOpen-source personal AI assistant & Agent Harness. Plans tasks, runs tools and skills, self-evolves with memory and knowledge. Multi-agent, multi-model, multi-channel. Lightweight, extensible, one-line install.
Languages
Trust signals
From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.
