Fellow

Fellow spent years as a meeting-management platform — agendas, one-on-one templates, action-item tracking — with AI notes bolted on the side. That is no longer the product. With the Fellow 5.0 release in December 2025 and a “Privacy First, Always” repositioning, the company rebuilt itself around AI meeting capture and an agent it calls an AI Chief of Staff. The marketing line is now “Your Secure AI Meeting Assistant,” and the AI is the core, not an accessory.

The clearest signal of that shift is how Fellow records. Botless desktop recording is a flagship capability: it pulls audio straight from your device, so it can capture in-person meetings, Slack huddles, or anything else playing through your speakers and mic without sending a bot into the call. There is also a Zoom-native mode that leans on Zoom’s own recording disclosures instead of a separate participant. Both modes show a visible recording indicator and let you pause and resume. For teams that found meeting bots awkward or non-compliant, this is the headline change.

Fellow keeps its cross-platform reach. With apps for Mac, Windows, iOS, Android, and web, it remains one of the more broadly accessible tools covered here. The integration ecosystem is deep — Zoom, Google Meet, Teams, Slack, Jira, Asana, Salesforce, HubSpot — and Fellow now layers a public API, Zapier automations, and an MCP server on top of those native connectors.

Key Features

The centerpiece is Ask Fellow, an agent that runs retrieval across every meeting you have access to and answers questions in plain language: what did we decide about the Q3 roadmap, who owns the migration, what changed since last week. It goes beyond answering, too — it can draft emails, generate docs and memos, book follow-ups, and push CRM updates. Around that agent sit AI meeting notes, summaries, and action items generated from each recording, plus transcription and AI notes in 92 languages with automatic language detection.

The MCP server is worth calling out. It exposes your meeting notes, transcripts, and action items as a secure, permissioned data source that external AI tools like Claude and ChatGPT can query directly, which makes Fellow a source of meeting context rather than a walled garden. The public API (launched August 2025) and Zapier automations extend that into custom workflows. Fellow’s older strengths — collaborative agendas, templates for recurring meetings and one-on-ones, an action-item tracker, and meeting analytics — are still present, but they now frame the AI rather than the other way around.

Pricing

Fellow’s free plan is effectively a trial. It allows up to 5 seats but only 5 AI notes and 5 AI recordings per user for the entire lifetime of the account, so it is enough to evaluate the product and little more. The Team plan (formerly Pro) is $7 per user per month billed annually, or $11 month-to-month, and covers up to 20 seats with 10 AI notes and 10 AI recordings per user each month plus API access. Business is $15 per user per month annually (or $23 monthly), supports up to 50 seats, and unlocks unlimited AI notes and recordings along with Salesforce and HubSpot CRM sync and keyword tracking. Enterprise runs $25 per user per month annually with a 10-user minimum and adds SSO, transcript redaction, AI-powered CRM updates, and org-wide analytics. Individuals can also take a Solo plan at $19 per user per month annually ($29 monthly) for unlimited AI notes and recordings.

The Team price is competitive, but note where “unlimited” actually starts: the cap of 10 AI notes and recordings per month on Team means heavy users will be pushed toward Business at more than double the seat cost. Budget around the AI limits, not just the headline number.

Privacy & Security

This is where Fellow has changed the most. The product is now SOC 2 Type II, GDPR, and HIPAA compliant, and it will sign Business Associate Agreements for regulated workflows — a genuinely privacy-forward posture rather than the table-stakes one the category usually offers. Fellow encrypts data at rest with AES-256, states plainly that it never trains its AI on customer data, and gives admins granular controls over who can record and what gets captured. Enterprise customers get transcript redaction on top of that.

A couple of caveats keep this honest. Fellow processes and stores your recordings in its cloud — meeting audio leaves your machine, with the same fundamental exposure that comes with any cloud meeting tool. Its infrastructure runs in a single AWS region (Canada Central), so despite the strong compliance story, this is not a tool that offers selectable multi-region data residency. For most teams the governance controls will be more than enough; for those with strict on-device or in-region mandates, the cloud architecture is still the architecture.

Best For

Fellow fits teams that want a privacy-conscious AI meeting assistant and value capturing meetings without a bot — particularly organizations in healthcare or other regulated fields that need HIPAA coverage and a no-training-on-data guarantee. It is well suited to companies that already live across Zoom, Meet, and Teams and want one assistant that can record in-person sessions too, then answer questions across all of them through Ask Fellow. The lingering agenda and one-on-one tooling makes it a natural pick for managers who still want meeting structure alongside the AI. It is a weaker fit for anyone who needs offline capture or fully on-device processing, since Fellow is cloud-only.

How Fellow Compares to Hedy

Fellow and Hedy have converged more than they used to. Both now offer bot-free capture — Fellow records directly from the desktop without a participant bot, and Hedy captures system audio the same way — so the old “bot versus botless” contrast no longer separates them. Both lean hard into privacy as a selling point, and both let you ask questions across your meeting history.

The durable difference is where the AI runs. Hedy processes end-to-end on-device: transcription, summaries, and real-time coaching all happen locally, and it keeps working without an internet connection. Fellow’s AI runs in its cloud — it pairs that with serious compliance credentials (HIPAA, SOC 2 Type II, a no-training pledge) and admin controls, but your audio is still processed and stored externally, and there is no offline mode. So the choice comes down to model: if you want strong cloud governance with botless capture, broad integrations, and an agent that can take actions and push CRM updates, Fellow is a strong option. If your requirement is that meeting audio and AI never leave the device — including live coaching during the conversation and full offline use — Hedy is built for that, and Fellow’s cloud architecture cannot match it.