Skip to content
Home About Services Portfolio Blog Contact Get Started
AI & Automation

Spellar 3.0 Review: An AI Meeting Assistant That Actually Remembers

Written by

MTD Technologies

Published
Read Time 8 min
Spellar AI meeting assistant summarizing a video call with automatic notes and cross-meeting memory

Every AI meeting assistant on the market promises to end note-taking forever. Most of them deliver the same disposable artefact: a summary you skim once and never open again. Spellar 3.0 takes a different bet — that the real value of an AI meeting companion is not the summary of one call, but memory across all of them. After reviewing its Product Hunt launch (May 2026), its pricing, its reviews, and what it actually does under the hood, here is our verdict: Spellar is the strongest pick for solo professionals and small teams who live in client calls, with one platform caveat you need to know before paying.

Spellar AI meeting assistant reviewing

What Spellar does as an AI meeting assistant

Spellar (from Spellar LTD) is a native Mac, iPhone, iPad and web app that records your meetings and turns them into structured notes. Its headline differentiator is architectural: unlike Otter.ai or Fireflies, which join your call as a visible bot attendee, Spellar records audio directly on your device. No “Spellar Bot has joined the meeting” announcement, no awkward silence while everyone stares at the robot.

Here is what it actually gives you:

  • Meeting notes and summaries with action points and a clear structure, generated automatically after each call — plus task extraction from the transcript.
  • Cross-meeting memory: ask what was decided in a thread of meetings, and Spellar answers from context across calls — 1:1s, follow-ups, and project reviews all stay connected instead of living in isolated summaries.
  • Real-time speaking feedback: live tips during calls, filler-word detection, pace and clarity feedback, and post-call notes on pronunciation and vocabulary. This grew out of the product’s origins as a speaking co-pilot.
  • 100+ language transcription, with on-device transcription on Apple hardware (server-side AI transcription is optional and off by default).
  • Model choice on Pro: switch between GPT, Claude, Gemini and Perplexity per meeting, per template, or per question — and you can connect your own API keys if you want processing fully under your control.
  • 20+ integrations: one-tap exports to Notion, Google Docs, Linear, Slack, Jira, Asana, Todoist, Craft and Miro, plus a developer API, webhooks, and MCP server access on Pro for wiring meeting data into your own pipelines.

Spellar pricing: what it actually costs

We checked pricing from multiple directory listings because Spellar’s pricing differs between its web plans and its App Store plans — a common source of confusion, so here are both:

  • Pro (web): US $11.99 per month when billed annually (US $143.88 per year). Includes unlimited AI recordings, all premium AI models, on-device transcription, 18–20+ integrations, 14-day audio retention, API and webhooks, and MCP server access. A 14-day money-back guarantee applies.
  • App Store plans: Monthly at US $12.99, Annual at US $95.99; Pro+ tiers at US $15.99 per month or US $129.99 per year. Recent release notes confirm that free integrations (Notion, Google Docs, Todoist) no longer show a Pro paywall on Mac — so the free tier is usable for basic capture.
  • Spellar Teams: shared templates, folders and notes, centralised billing and team integrations, starting at five seats with pricing on request via sales.

There is no permanent free tier on the web Pro path — treat the guarantee as your trial. At $11.99 per month annually, it sits in the same band as Fathom and Otter’s paid tiers, so the buying decision comes down to features, not price.

Spellar AI meeting assistant cross-meeting memory concept

Who Spellar is actually for

The compounding-memory angle is not marketing fluff — it maps to a real workflow. If you run weekly 1:1s, manage a portfolio of client accounts, or do discovery calls that turn into proposals, the questions you actually ask are things like “what did we agree to change after last month’s call?” and “which commitments are still open across all three stakeholders?” A bot that summarises each call in isolation cannot answer those. Spellar’s cross-meeting memory can.

It is a particularly good fit for:

  • Consultants and freelancers who turn discovery calls into proposals and need to track what each client said, when.
  • Non-native English speakers doing business in English — the speaking coach (filler words, pace, vocabulary suggestions) is a genuinely rare feature in this category.
  • Small teams on Apple hardware that want meeting notes flowing into Notion or Linear without a bot announcing itself on every call.
  • Privacy-conscious users who want on-device transcription and the option to bring their own API keys.
Spellar AI meeting assistant real-time speaking coach

Watch-outs and honest limitations

No review is useful without the downsides. Here are the ones that matter:

  • Apple-only. Spellar runs on Mac, iPhone, iPad and the web. There is no native Windows or Android app. If your team is mixed-platform, the web app is your only fallback, and that weakens the no-bot-recording story on PCs.
  • The bot-free approach has a trade-off. Recording on-device means it captures your side perfectly and relies on device audio for everyone else — several user reviews flag the occasional recording hiccup (one reports issues with Zoom and AirPods on Mac). Bot joiners are awkward, but they record server-side with near-perfect reliability. Pick your poison.
  • Pricing is fragmented. Web Pro, App Store Pro, and App Store Pro+ all exist at different price points. Decide which platform you will pay through before you subscribe, and check what the free tier still includes — it has improved recently (free integrations unlocked on Mac).
  • Young product, small footprint. Spellar’s team and user base are small compared to Otter or Fireflies (its Product Hunt rating is a stellar 4.96/5, but from only 23 reviews). That is not a dealbreaker, but it is why we checked its launch history — the team has shipped consistently since 2023, which is a good sign.
  • Privacy posture is strong but not magic. On-device transcription and optional local processing are real differentiators, but if you enable cloud models for summaries, audio does leave the device. Configure audio auto-deletion to match your retention policy — the app supports automatic deletion after a chosen duration.

How Spellar compares to the alternatives

The AI meeting assistant space splits into two camps. Bot joiners (Otter.ai, Fireflies.ai, Fathom) send a bot into your call, record server-side, and work on any platform — at the cost of the awkward “recording bot has joined” moment that changes how people talk. Bot-free recorders (Spellar, Granola) run on your device and keep meetings feeling human — at the cost of platform lock-in and device-side reliability.

Within the bot-free camp, Granola is Spellar’s closest rival: similar no-bot recording, strong summaries, heavy on the “stay present in the meeting” pitch. Spellar’s edges are cross-meeting memory (asking questions across your meeting history), the speaking coach, and model choice. Granola’s edge is maturity and a bigger user base. If your main pain is “I need notes from every call,” either works. If your pain is “I need to remember what was decided across six weeks of calls,” Spellar’s memory model is the more interesting bet. We cover more tools like these in our AI Tools Directory.

From our experience: making meeting AI actually pay off

We have wired meeting-note tooling into client workflows for years, including automated n8n workflows that push call summaries into CRMs and project boards. The pattern we see repeatedly: the notes themselves are not the bottleneck — distribution is. A summary that sits in the recorder’s app helps one person. A summary that lands in your Notion database, creates Linear tasks with owners, and pings Slack creates organisational memory.

That is why Spellar’s integrations and MCP server access matter more than its transcription accuracy (which is table stakes in 2026). When evaluating any AI meeting assistant, run it for two weeks against this checklist: do action items reach your task system without copy-paste? Can you search a decision across meetings, not just within one? Does it work on every device your team actually uses? Spellar scores well on the first two and poorly on the third if you are not an Apple shop. Be honest about which of those matters most for your business — and talk to us if you want help plumbing meeting intelligence into your existing stack.

Frequently asked questions

Is Spellar’s AI meeting assistant free?

Spellar has a free tier with basic recording and notes — recent updates removed the Pro paywall from free integrations like Notion, Google Docs and Todoist on Mac. The full Pro plan is US $11.99 per month billed annually, with a 14-day money-back guarantee instead of a free trial.

Does Spellar work with Zoom and Google Meet?

Yes. Spellar records device audio, so it works with Zoom, Google Meet, Microsoft Teams, Webex and in-person conversations captured on iPhone or iPad. No bot joins your call — the trade-off is that recording quality depends on your device and audio setup rather than a server-side bot.

What makes Spellar 3.0 different from Otter.ai or Fireflies?

Three things: cross-meeting memory (it answers questions across your meeting history, not just within one call), a real-time speaking coach with filler-word and pace feedback, and your choice of AI model (GPT, Claude, Gemini or Perplexity) — plus no bot joiner, since it records on your device.

Is Spellar available on Windows or Android?

Not natively. Spellar runs on Mac, iPhone, iPad and the web — the web app is the only option on Windows or Android. Mixed-platform teams should test the web experience carefully before committing, or consider a bot-based alternative that is platform-agnostic.

Is Spellar private enough for confidential business meetings?

It is stronger than most competitors: on-device transcription, optional server-side processing (off by default), the ability to connect your own API keys, and automatic audio deletion. For highly regulated industries, review the retention settings and your own compliance requirements before rolling it out team-wide.

Bottom line: Spellar 3.0 is a genuinely differentiated AI meeting assistant. Its cross-meeting memory turns months of calls into a searchable record, the speaking coach is unique in the category, and the privacy story is credible. The Apple-only reality and device-side recording quirks are the real costs. If you run your business on a Mac and live in client meetings, it is worth the Pro plan. If your team is mixed-platform, test the web app first — or ask us to help you evaluate alternatives against your actual workflow.