Fellow.ai MCP
Fellow.ai MCP lets agents access authorized meeting notes, transcripts, calendar events, channels, action items, and agenda workflows.
Verdict
Common use cases
- Summarize action items from last sprint retro
- Pull agenda topics from leadership sync
- Track recurring standup commitments over time
- Draft follow-up email from client meeting notes
- Compare decisions across product planning sessions
Integration
- Vendor
- Fellow.ai MCP
- Category
- other
- Auth
- OAUTH2
- Composio slug
fellow_ai_mcp
Tools
Setup
Setup guide
- 11. Open your Switchy workspace settings and navigate to the Integrations tab. 2. Find Fellow.ai MCP in the directory and click Connect. 3. You'll redirect to Fellow's OAuth consent screen — sign in with your Fellow account if prompted. 4. Grant read access to meetings, agendas, and action items (Fellow will list the specific scopes). 5. After approving, you'll return to Switchy with a green Connected badge next to Fellow.ai MCP. 6. Open any Project and type '@Fellow' in the chat to confirm the connection — ask it to list your recent meetings or summarize action items from the last week. 7. If the MCP responds with meeting titles or agenda snippets, setup succeeded.
What teammates see: by default, memories from Fellow.ai MCP are scoped to the Project (PROJECT visibility) - you can mark any memory PRIVATE or share it ORG-wide.
Works well with
Top models
Compatibility data appears once enough Projects have used this MCP together with a given model.
How Switchy teams use it
Starter prompts
Weekly Action Item Rollup
@Fellow list all open action items assigned to anyone on my team from meetings in the last 7 days, grouped by person.
Standup Prep Summary
@Fellow summarize the key discussion points and any action items from yesterday's standup meeting.
Retro Themes Analysis
@Fellow review the last three sprint retrospectives and highlight any themes or issues that came up more than once.
Meeting Decision Log
@Fellow pull all decisions recorded in our product roadmap meetings over the last month and format them as a bullet list.
Client Meeting Follow-Up
@Fellow draft a follow-up email for our client based on today's meeting notes, including next steps and any commitments we made.
Example outputs
Illustrative - representative of the model's voice and quality, not literal recordings.
@Fellow.ai show me all action items assigned to me from meetings this week
Based on your Fellow.ai account, here are 7 action items assigned to you this week: (1) Review Q1 budget proposal from Monday's finance sync, due Friday. (2) Send onboarding doc to new hire from Tuesday's team standup, due today. (3) Schedule follow-up with design team from Wednesday's product review, no due date set. (4) Approve marketing copy from Thursday's campaign planning, due next Monday. The remaining 3 items are marked complete. All items pulled from 4 different meetings across your calendar.
This example assumes the MCP can query meeting notes and extract structured action items. Requires OAuth scope covering meeting read access. If your Fellow.ai workspace uses custom fields or tags for action items, the MCP's parsing may vary. No side-effects — read-only query.
@Fellow.ai create a recurring 1-on-1 agenda template for my direct reports with sections for wins, blockers, and career development
Created a new recurring agenda template in your Fellow.ai workspace titled '1-on-1: Direct Report Check-in'. Template includes three sections: (1) Wins & Progress — space for recent accomplishments, (2) Blockers & Support Needed — current obstacles requiring manager input, (3) Career Development — long-term growth topics and skill-building. This template is now available when scheduling 1-on-1 meetings with any team member. You can edit section names or add talking points directly in Fellow.ai.
Demonstrates the MCP's ability to scaffold meeting structures. Requires OAuth scope for creating templates or agendas. The template persists in your Fellow.ai account — this is a write operation with lasting side-effects. Confirm the template matches your team's workflow before using it widely.
@Fellow.ai summarize the key decisions and open questions from our last three product roadmap meetings
Across your last three product roadmap meetings (Jan 15, Jan 22, Jan 29), the team decided to: (1) prioritize mobile app performance over new features in Q1, (2) delay the analytics dashboard redesign to Q2, (3) allocate two engineers to address technical debt. Open questions still unresolved: Should we support offline mode for the mobile app? What's the budget impact of the Q2 delay? Who owns the customer communication plan for the feature shift? These themes appeared in meeting notes tagged 'roadmap' and 'product-strategy'.
Combines the MCP's meeting data access with the AI's synthesis capability. Useful for catching up after missing meetings or preparing for follow-ups. Accuracy depends on how thoroughly your team documents decisions in Fellow.ai — incomplete notes yield incomplete summaries. Read-only operation.
Use-case deep-dives
When meeting prep needs to pull from multiple sources
A 6-person leadership team runs a weekly sync where each person reports on their area. Without Fellow, someone manually copies updates from Slack threads, Linear, and last week's notes into a Google Doc 30 minutes before the call. With the Fellow MCP, Switchy can draft the agenda by pulling action items from the last meeting, open blockers from project tools, and recent decisions from chat—all in one prompt. The OAuth2 flow means each leader authenticates once and Switchy respects their individual permissions. This works when your team already uses Fellow for meeting notes and you want to stop the pre-meeting copy-paste ritual. If your team doesn't use Fellow yet, adding it just for MCP access is overkill—stick with direct integrations to your project tracker instead.
Pulling meeting history into onboarding context
A 12-person startup hires a new engineer who needs to understand how the team makes decisions. The onboarding buddy currently screenshots meeting notes and pastes them into Notion, losing context and wasting 90 minutes per hire. The Fellow MCP lets Switchy query past meetings by topic—'show me all architecture decisions from Q1'—and generate a structured summary with links back to the original notes. OAuth2 means the new hire only sees meetings they're invited to, which matters for confidential leadership discussions. This is the right call if your team already documents decisions in Fellow and hires frequently enough that manual assembly hurts. If you only onboard twice a year, the manual process is fine and you don't need the MCP overhead.
When you need to surface decisions across siloed meetings
A product manager is kicking off a feature that touches engineering, design, and support. Each function has been discussing it in their own weekly meetings for a month, and no one has a full picture. The PM currently asks each lead to summarize their team's stance in Slack, then stitches it together manually. With the Fellow MCP, Switchy can query all three teams' meeting notes for mentions of the feature name and produce a unified timeline of decisions, open questions, and owners. This works because Fellow already holds the canonical record and OAuth2 ensures Switchy only reads meetings the PM has access to. The threshold: if your cross-functional projects happen less than once a quarter, just ask people directly. If they happen weekly and meeting notes are scattered, the MCP pays for itself in hours saved.
Frequently asked
What does the Fellow.ai MCP do in Switchy?
It connects your Fellow.ai meeting workspace to Switchy's AI agents so they can reference meeting notes, action items, and agendas when answering questions or drafting documents. Your team gets context-aware responses that pull from actual meeting discussions instead of guessing what was decided. The MCP uses OAuth2 so you authorise once and Switchy handles token refresh automatically.
Do I need admin rights in Fellow.ai to connect this MCP?
No, any Fellow.ai user can connect the MCP through OAuth2. You'll authorise Switchy to read your accessible meetings and notes — the scopes match your existing Fellow.ai permissions. If your workspace restricts certain meetings to specific teams, the MCP respects those boundaries and won't surface content you can't already see in Fellow.ai.
Can the MCP create new meetings or edit action items in Fellow.ai?
Not yet. The current integration is read-only, so Switchy's agents can reference meeting content but won't write back to Fellow.ai. If you need to update an action item or add notes, you'll still open Fellow.ai directly. This keeps your meeting records safe from accidental AI edits while giving agents the context they need.
Why use this instead of just copying meeting notes into Switchy?
The MCP pulls fresh data every time an agent needs it, so you're never working from stale notes. It also surfaces related meetings automatically — if you ask about a project decision, the agent can check three months of relevant standups without you remembering which notes to paste. Manual copying breaks the moment someone updates an action item in Fellow.ai.
Who on the team should connect the Fellow.ai MCP?
Anyone who regularly references meeting notes in their work. Each person connects their own Fellow.ai account, so agents see only the meetings that user can access. If your ops lead runs retrospectives and your PM owns planning meetings, both should connect — then agents can pull context from either person's meetings depending on who's asking the question.