Speechmark

Speechmark is a macOS meeting notes app that records, transcribes, and summarizes meetings entirely on your Mac. From the menu bar, one click starts a recording that captures your microphone and the meeting audio together, and the app handles the audio routing for you — there is no browser extension and no bot that joins the call. Speakers are separated automatically and labelled the first time you confirm a name, so every transcript afterwards reads like a script. The app then writes the note you would have written: decisions, action items, and a short recap, all in plain editorial prose. It is built for people whose meetings matter — product managers, legal and compliance teams, consultants, founders, and executives — and it needs no account, no login, and no sign-up. Most meeting recorders are cloud-first tools. Otter.ai and Fireflies upload your audio to their servers and charge per seat, while Granola transcribes on-device but syncs your notes to its cloud and, unless you opt out, uses anonymized meeting data to improve its AI models. Speechmark positions itself as the private alternative to these cloud recorders. Its comparison table states that, unlike Otter.ai and Fireflies, Speechmark keeps audio on your Mac, transcribes on-device, requires no account or login, sends no bot into your call, uses one-time pricing, and lets you use your own AI key. The company notes that competitor data reflects those products' cloud-first defaults and that features may vary by plan. For people who discuss sensitive material — legal matters, compliance reviews, product strategy — the difference between audio that stays local and audio that is uploaded is the whole point. Capture in Speechmark lives in the Mac menu bar. One click begins recording, and the app captures your microphone and the meeting audio together while handling the routing for you. There is no browser extension to install and no bot that dials into the call, so other participants see no recording bot. Speechmark also handles speaker attribution: speakers are separated automatically, and they are labelled the first time you confirm a name. After that confirmation, every future transcript reads like a script, with each line attributed to the person who said it. This matters because speaker attribution turns a raw transcript into something you can scan and quote. In the Claude example shown on the site, an answer cites follow-up items attributed to specific people such as “you” and “Dana”, with the meeting each item came from listed beneath it. The writing step produces the note you would have written rather than a wall of bullet points. Speechmark generates decisions, action items, and a short recap, all in plain editorial prose. You can choose the model that fits your privacy bar: Apple Intelligence, a local Ollama model, or your own OpenAI or Anthropic key. That choice means you can keep summarization fully on-device with Apple Intelligence or a local Ollama model, or send transcript text only — never audio — to a cloud model using your own key. Every line of the generated note is clickable: clicking any line jumps to the exact moment in the meeting when it was said. That link between the summary and the source keeps you close to what was actually agreed, so you can verify a decision or check the context around an action item without listening to the entire recording again. Speechmark also connects to Claude. A section labelled “New · Works with Claude” explains that connecting Speechmark to Claude Desktop turns your transcript folder into a memory your own AI can reason over, privately and on your device. The setup is one click: Settings → Assistant → Add to Claude Desktop installs it as a Claude extension, with no config files, no terminal, and no API key. Once connected, Claude answers from your own notes — decisions, action items, attendees — and cites the meeting each answer came from. The connector reads your notes locally and makes no network calls of its own, so you decide what any question surfaces. The site illustrates this with a sample question, “What did we decide about pricing across my Acme calls?”, answered across three Acme calls with a conclusion about annual billing at a 15% discount, two unresolved follow-ups, and the source meetings listed. The Product Hunt description additionally notes that Speechmark integrates well with Claude Code using an MCP server. Speechmark's overall approach is privacy by architecture. Audio never leaves your Mac: recording, transcription, and note generation all happen on the device. Transcription runs locally on Apple silicon, and notes are generated on-device by default. The site's diagram shows that the only thing that can ever leave is transcript text, and only if you opt in by choosing a cloud model — audio never arrives in the cloud, is never stored, and is never streamed. The privacy section lists the guarantees plainly: recordings and transcripts stay on your Mac, transcription runs on-device, no bot is dialed into your call, there is no account, login, or sign-up, usage stats are opt-in and off by default, and a cloud model sends text only, never audio. As the FAQ puts it, if you pick a cloud AI model for the summary (OpenAI or Anthropic, with your own key), only the transcript text is sent, never the audio, and you can also stay fully on-device with Apple Foundation Models or a local Ollama model. The benefits follow directly from that design. An early-access user quoted on the site, Mira K., director of product, says she used to leave meetings with three pages of bullet points and no idea what had actually been agreed, and now leaves with a paragraph that reads like a memo from someone who took it seriously. Because Speechmark tracks who said what and when, sprint retros, PRD updates, and stakeholder readouts practically write themselves, and you can stay in the conversation instead of racing to keep notes. The emphasis on decisions and action items rather than an exhaustive bullet list is aimed at teams that will not read a wall of bullet points. Clickable lines that jump to the exact moment give you a way to check what you are reading against what was actually said, and the absence of a bot means participants are not confronted with a recorder joining the call. Speechmark's stated use cases centre on meeting-heavy roles. The product-manager description covers capturing the decisions and action items rather than a wall of bullet points, and letting sprint retros, PRD updates, and stakeholder readouts write themselves. The “built for people whose meetings matter” section lists four groups: product managers, legal and compliance, consultants, and founders and executives. Beyond note-taking, the Claude integration supports a research workflow: asking questions across your meeting history — for example, what was decided about pricing across several calls with one client — and receiving an answer grounded in your own notes that cites the meeting each answer came from, including unresolved follow-ups and the person responsible for each. Speechmark runs on macOS 14.2 (Sonoma) or later on an Apple silicon Mac. The download is listed as v1.0.0 and 13 MB. There is no account, credit card, or waitlist to try it: you can download and run it. Pricing is a one-time purchase rather than a subscription, and a single purchase covers up to three of your Macs. The $49 founding-customer price applies to the first 200 customers; the price is then $79. Optional model integrations include Apple Intelligence, a local Ollama model, and your own OpenAI or Anthropic key, plus the one-click Claude Desktop connector. Product Hunt lists the product under the topics Mac, Notes, and Meetings. Release notes are the only thing the optional email signup promises, with no marketing and no list sharing, and a feedback email address is provided for bugs and requests. Speechmark's core value proposition is straightforward: it turns meetings into clean, editorial notes — decisions, action items, and speaker-attributed transcripts — while keeping audio, transcripts, and notes on your Mac. No bot joins the call, no account is needed, and you pay once instead of subscribing. For anyone whose meetings carry sensitive content, or who simply wants what was agreed to be captured without effort, Speechmark offers a private, on-device alternative to cloud recorders, with an optional Claude integration that lets you ask your own meeting history questions without the notes leaving your machine.