# How to integrate Wispr Flow MCP with Atomic Agent ```json { "title": "How to integrate Wispr Flow MCP with Atomic Agent", "toolkit": "Wispr Flow MCP", "toolkit_slug": "wispr_flow_mcp", "framework": "Atomic Agent", "framework_slug": "atomic-agent", "url": "https://composio.dev/toolkits/wispr_flow_mcp/framework/atomic-agent", "markdown_url": "https://composio.dev/toolkits/wispr_flow_mcp/framework/atomic-agent.md", "updated_at": "2026-09-30T05:38:17.783Z" } ``` ## Introduction Atomic Agent is an open-source AI agent that runs on your machine and supports local models through llama.cpp. Connect it to 1,500+ apps through Composio, including Wispr Flow. This guide uses Atomic Agent's built-in Composio integration. You need Atomic Agent v0.5.6 or later, your own Composio API key, and a Wispr Flow account. See the [Atomic Agent documentation](https://atomicagent.io/docs/?cta_placement=atomic-agent-docs) for installation instructions. ## Also integrate Wispr Flow MCP with - [ChatGPT](https://composio.dev/toolkits/wispr_flow_mcp/framework/chatgpt) - [Claude Cowork](https://composio.dev/toolkits/wispr_flow_mcp/framework/claude-cowork) - [Hermes](https://composio.dev/toolkits/wispr_flow_mcp/framework/hermes-agent) ## TL;DR ### How Atomic Agent connects to Wispr Flow Atomic Agent connects to Composio's Tool Router over Streamable HTTP MCP. It discovers the tools needed for your request, and Composio handles account authentication. Execution and account-connection tools follow Atomic Agent's configured approval policy. Review any approval prompt before allowing an action on your connected account. ## Connect Wispr Flow MCP to Atomic Agent ### Connect Wispr Flow to Atomic Agent - Sign up or sign in to the [Composio dashboard](https://dashboard.composio.dev/?cta_placement=atomic-agent-api-key). Copy your API key. - In Atomic Agent, open the Integrations tab and select Composio. Select API key, press e, paste the key, and press Enter. The tools become available immediately, without a restart. - Ask Atomic Agent to do something in Wispr Flow. It searches Composio for the right tool. - On first use, open the sign-in link returned in the chat. Connect your Wispr Flow account, then return to Atomic Agent to continue. ### Headless setup Add the following line to ~/.atomic-agent/.env, replacing YOUR_COMPOSIO_API_KEY with your key. Start Atomic Agent after saving the file. ```bash COMPOSIO_API_KEY=YOUR_COMPOSIO_API_KEY ``` ## What is the Wispr Flow MCP server, and what's possible with it? The Wispr Flow MCP server is an implementation of the Model Context Protocol that connects your AI agent and assistants like Claude, Cursor, etc directly to your Wispr Flow account. It provides structured and secure access so your agent can perform Wispr Flow operations on your behalf. ## Supported Tools | Tool slug | Name | Description | |---|---|---| | `WISPR_FLOW_MCP_GET_ACCOUNT_INFO` | Get account info | Get the authenticated user's own identity — their name and any display-name aliases they appear under on their calendar. Call this first for first-person questions ('what did I commit to', 'my meetings', 'who did I meet with', 'did I owe him') so you know which person the user is among a meeting's attendees. | | `WISPR_FLOW_MCP_GET_CALENDAR_EVENT` | Get calendar event | Get a single calendar event by its calendar_id (the id returned by search_calendar_events or resolve_calendar_link). Returns title, time, attendees, description, and conference URL. Every timestamp in the response is UTC (ISO 8601, ending in 'Z'). Convert times to the user's local time zone before presenting them — do not show the raw UTC time to the user. | | `WISPR_FLOW_MCP_GET_MEETING` | Get meeting | Get a bounded range of the markdown notes, the complete derived markdown summary, the complete attendee list without email addresses, and (when requested and available) a bounded range of the recorded transcript of a specific meeting. Use after search_meetings to answer anything about what was discussed, decided, the action items, or who-said-what. Returns start/end times for when the meeting happened. The raw Lexical notes JSON is intentionally omitted; use `content` for the paginated markdown note body and `summary` for the Flow Summary body, whose next-steps section holds the meeting's action items. Prefer view_transcript={} and synthesize your answer from the transcript itself: the notes are an auto-generated summary that can drop details, while the transcript is the verbatim source of truth. Omit view_transcript only when you've judged the notes alone are enough (e.g. a quick high-level recap). Transcript responses are character-bounded; when truncated, use the continuation offset in the marker to request the next range. Content is also character-bounded and returns its first range by default; use view_content with the continuation offset to read more. The summary is always returned in full. share_link is the meeting's notes page in Wispr Flow. Give it to the user whenever they ask for a link. Who else can open it depends on that meeting's sharing setting, and this server cannot read or change that setting. If the user wants to change who a link works for, tell them to open the Wispr Flow app, go to the Notetaker tab, click the menu to the right of the meeting, and click Share — that opens the sharing settings for that meeting. A meeting with no share_link has no notes page yet — say so rather than guessing a URL. Every timestamp in the response is UTC (ISO 8601, ending in 'Z'). Convert times to the user's local time zone before presenting them — do not show the raw UTC time to the user. | | `WISPR_FLOW_MCP_GET_MEETING_BY_CALENDAR_ID` | Get meeting by calendar | Get the recorded meeting — a bounded range of its markdown notes, the complete derived markdown summary, and (when requested and available) a bounded transcript range — captured for a calendar event, by its calendar_id (from search_calendar_events or resolve_calendar_link). Use this to answer 'what was discussed / decided / the action items' for a meeting on the user's calendar. Returns start/end times for when the meeting happened. The raw Lexical notes JSON is intentionally omitted; use paginated `content` and complete `summary`, whose next-steps section holds the meeting's action items. Prefer view_transcript={} and synthesize from the transcript itself: the notes are an auto-generated summary that can drop or distort details, while the transcript is the verbatim source of truth. Omit view_transcript only when you've judged the notes alone are enough. When truncated, use the continuation offset in the marker to request the next range. Content returns its first bounded range by default; use view_content with its continuation offset to read more. Errors if no recording is linked to that event. share_link is the meeting's notes page in Wispr Flow. Give it to the user whenever they ask for a link. Who else can open it depends on that meeting's sharing setting, and this server cannot read or change that setting. If the user wants to change who a link works for, tell them to open the Wispr Flow app, go to the Notetaker tab, click the menu to the right of the meeting, and click Share — that opens the sharing settings for that meeting. A meeting with no share_link has no notes page yet — say so rather than guessing a URL. Every timestamp in the response is UTC (ISO 8601, ending in 'Z'). Convert times to the user's local time zone before presenting them — do not show the raw UTC time to the user. | | `WISPR_FLOW_MCP_GET_MEETING_PARTICIPANTS_ENRICHED` | Get meeting participants enriched | Who was in a meeting and which one of them is the user. Returns the user's own name/email/domain plus every participant (speaker-map people and calendar attendees) with name, contact_id, emails, whether they spoke, organizer, and relationship to the user (self \| colleague \| external \| unknown, by email domain). Omits the caller, mention-origin people, and Google room/equipment attendees. Calendar RSVP is not included. When listed, get_participant_contacts takes contact_ids from this result for LinkedIn, job title, and company. | | `WISPR_FLOW_MCP_GET_SCRATCHPAD_NOTE` | Get scratchpad note | Get a bounded range of a specific note's normalized text. The first range is returned by default. Call after search_scratchpad_notes when the user wants the note's content; when truncated, repeat with view_content.start_char set to the exact continuation offset in the marker. | | `WISPR_FLOW_MCP_GET_UPCOMING_MEETING` | Get upcoming meeting | Get a single upcoming calendar event by calendar_id, with its latest pre-read when one is available. Refuses past events — use get_meeting_by_calendar_id for the recorded notes of a past event, or get_calendar_event for generic event metadata. Pass include_preread=false to skip the pre-read fetch. Every timestamp in the response is UTC (ISO 8601, ending in 'Z'). Convert times to the user's local time zone before presenting them — do not show the raw UTC time to the user. | | `WISPR_FLOW_MCP_LIST_MEETING_SERIES` | List meeting series | Given a meeting_id (from search_meetings) for a recurring meeting, list every recorded occurrence of that recurring series — newest first, including the meeting you passed in. Use to answer 'what did we discuss last time this met', 'what's recurred across these syncs', or to catch up before today's instance. Errors if the meeting isn't linked to a recurring calendar series. Each occurrence has the same shape as a search_meetings result (call get_meeting for full notes/transcript). Results are paginated, newest first. If the response's has_more is true there are more matches: call again with cursor set to the returned next_cursor (copied verbatim) to continue. When more remain, a 'more' field spells out how many and how to continue. To answer 'all'/'every'/'ever' questions, keep paging until has_more is false. Paging is capped at 1000 results per query; truncated=true means more matches exist beyond the cap — narrow the query. Every timestamp in the response is UTC (ISO 8601, ending in 'Z'). Convert times to the user's local time zone before presenting them — do not show the raw UTC time to the user. | | `WISPR_FLOW_MCP_LIST_UPCOMING_MEETINGS` | List upcoming meetings | List the user's upcoming calendar events in the next window_hours (default 24, max 168), soonest first, with an attendee preview (up to 5 structured attendees plus attendee_count) and each event's latest pre-read attached when one has been generated. Use to answer 'what meetings are coming up' or 'prep me for my next meeting'. For today or another whole local day, use search_calendar_events with since/until. Events without a generated pre-read return preread: null — pre-reads are produced on-demand by the desktop app, not by this tool. Results are paginated, newest first. If the response's has_more is true there are more matches: call again with cursor set to the returned next_cursor (copied verbatim) to continue. When more remain, a 'more' field spells out how many and how to continue. To answer 'all'/'every'/'ever' questions, keep paging until has_more is false. Paging is capped at 1000 results per query; truncated=true means more matches exist beyond the cap — narrow the query. Every timestamp in the response is UTC (ISO 8601, ending in 'Z'). Convert times to the user's local time zone before presenting them — do not show the raw UTC time to the user. | | `WISPR_FLOW_MCP_RESOLVE_CALENDAR_LINK` | Resolve calendar link | Resolve a calendar/meeting URL the user pasted (a Google Calendar event link, or a Meet/Zoom conference link) to the matching calendar event — title, time, attendees, conference URL, and its calendar_id. Use this whenever the user pastes such a link instead of guessing at its contents. To then pull the event again later, use get_calendar_event with the returned calendar_id. Every timestamp in the response is UTC (ISO 8601, ending in 'Z'). Convert times to the user's local time zone before presenting them — do not show the raw UTC time to the user. | | `WISPR_FLOW_MCP_RESOLVE_SHARE_LINK` | Resolve share link | Resolve a Wispr Flow shared-notes link the user pasted (https://notes.wisprflow.ai/shared/…) to the shared meeting note: its title, Flow Summary (whose next-steps section holds the meeting's action items), owner, and the caller's role. Use this whenever the user pastes such a link instead of guessing at its contents — shared notes are usually other people's meetings, which search_meetings/get_meeting (the user's own meetings) will never find. The owner also receives their user-authored My Notes. Set view_transcript={} when the user needs source-level detail: a bounded transcript range is returned only when the caller owns the note or is an explicitly invited recipient AND the owner's organization shares transcripts. Domain/team access, anyone-with-link access, or a notes-only organization returns transcript=null plus a transcript_access explanation when requested. When truncated, use the continuation offset in the marker to request the next range. Errors if the link is not a share link or no shared note is accessible for it. Every timestamp in the response is UTC (ISO 8601, ending in 'Z'). Convert times to the user's local time zone before presenting them — do not show the raw UTC time to the user. | | `WISPR_FLOW_MCP_SEARCH_CALENDAR_EVENTS` | Search calendar events | Search or list the user's Google Calendar events. Use to find meetings by who's on them (attendee_emails — the precise way to answer 'meetings with alex@acme.com', or 'with both alex and bob' via attendee_match=all), by keyword in the title/description, or within a date window. For a whole local calendar day (including earlier events), use since and until. Returns each event's title, time, conference URL, and an attendee preview (up to 5 structured attendees plus attendee_count). Use get_calendar_event for the complete attendee list. Note: these are calendar events, not the recorded meeting notes. Results are paginated, newest first. If the response's has_more is true there are more matches: call again with cursor set to the returned next_cursor (copied verbatim) to continue. When more remain, a 'more' field spells out how many and how to continue. To answer 'all'/'every'/'ever' questions, keep paging until has_more is false. Paging is capped at 1000 results per query; truncated=true means more matches exist beyond the cap — narrow the query. Every timestamp in the response is UTC (ISO 8601, ending in 'Z'). Convert times to the user's local time zone before presenting them — do not show the raw UTC time to the user. | | `WISPR_FLOW_MCP_SEARCH_MEETINGS` | Search meetings | Search or list the user's meetings captured by Wispr Flow Meeting Recorder. Use when the user asks about their meetings, calls, standups, 1:1s, syncs, or anything discussed in a recorded conversation. Returns the most recently modified meetings when no query is given. For meetings on a particular day, pass since/until (when the meeting happened, half-open). Filter by who was there with attendee_emails (matched against the meeting's linked calendar event) — the precise way to answer 'my meetings with alex@acme.com'. Each result lists up to 5 attendees (more are summarized — call get_meeting for the full list), start/end times, and a has_transcript flag; when true, follow up with get_meeting and view_transcript={} and answer from the transcript rather than the lossy notes summary. share_link is the meeting's notes page in Wispr Flow. Give it to the user whenever they ask for a link. Who else can open it depends on that meeting's sharing setting, and this server cannot read or change that setting. If the user wants to change who a link works for, tell them to open the Wispr Flow app, go to the Notetaker tab, click the menu to the right of the meeting, and click Share — that opens the sharing settings for that meeting. A meeting with no share_link has no notes page yet — say so rather than guessing a URL. Results are paginated, newest first. If the response's has_more is true there are more matches: call again with cursor set to the returned next_cursor (copied verbatim) to continue. When more remain, a 'more' field spells out how many and how to continue. To answer 'all'/'every'/'ever' questions, keep paging until has_more is false. Paging is capped at 1000 results per query; truncated=true means more matches exist beyond the cap — narrow the query. Every timestamp in the response is UTC (ISO 8601, ending in 'Z'). Convert times to the user's local time zone before presenting them — do not show the raw UTC time to the user. | | `WISPR_FLOW_MCP_SEARCH_SCRATCHPAD_NOTES` | Search scratchpad notes | Search or list the user's notes (Wispr Flow scratchpad). Use whenever the user asks about their notes, scratchpad, jottings, or anything they've written down. Returns the most recently modified notes when no query is given; case-insensitive substring match otherwise. Results are paginated, newest first. If the response's has_more is true there are more matches: call again with cursor set to the returned next_cursor (copied verbatim) to continue. When more remain, a 'more' field spells out how many and how to continue. To answer 'all'/'every'/'ever' questions, keep paging until has_more is false. Paging is capped at 1000 results per query; truncated=true means more matches exist beyond the cap — narrow the query. | ## Supported Triggers None listed. ## Troubleshooting ### Why is my API key rejected? Copy the key directly from the Composio dashboard. Atomic Agent requires ASCII characters because it sends the key in an HTTP header. Remove any extra characters introduced when copying. ### Why are Composio tools unavailable after saving a key? Check that you use Atomic Agent v0.5.6 or later and that Composio is enabled in your configuration. A "composio": { "enabled": false } setting disables the integration while keeping the saved key on disk. ### Does Composio have a free plan? The Hobby plan includes 100,000 tool calls per month. Composio-managed OAuth apps have a lower included allowance. See [Composio pricing](/pricing) for the current limits that apply to your connection. ## Complete Code None listed. ## Conclusion ### Use Wispr Flow from Atomic Agent With your account connected, ask Atomic Agent to work with Wispr Flow in natural language. You can connect more apps through the same Composio integration as you need them. Visit the [Atomic Agent website](https://atomicagent.io/?cta_placement=atomic-agent-site) or inspect the [Atomic Agent source code](https://github.com/AtomicBot-ai/atomic-agent?cta_placement=atomic-agent-repo). ## How to build Wispr Flow MCP Agent with another framework - [ChatGPT](https://composio.dev/toolkits/wispr_flow_mcp/framework/chatgpt) - [Claude Cowork](https://composio.dev/toolkits/wispr_flow_mcp/framework/claude-cowork) - [Hermes](https://composio.dev/toolkits/wispr_flow_mcp/framework/hermes-agent) ## Related Toolkits - [Google Sheets](https://composio.dev/toolkits/googlesheets) - Google Sheets is a cloud-based spreadsheet tool for real-time collaboration and data analysis. It lets teams work together from anywhere, updating information instantly. - [Notion](https://composio.dev/toolkits/notion) - Notion is a collaborative workspace for notes, docs, wikis, and tasks. It streamlines team knowledge, project tracking, and workflow customization in one place. - [Airtable](https://composio.dev/toolkits/airtable) - Airtable combines the flexibility of spreadsheets with the power of a database for easy project and data management. Teams use Airtable to organize, track, and collaborate with custom views and automations. - [Asana](https://composio.dev/toolkits/asana) - Asana is a collaborative work management platform for teams to organize and track projects. It streamlines teamwork, boosts productivity, and keeps everyone aligned on goals. - [Google Tasks](https://composio.dev/toolkits/googletasks) - Google Tasks is a to-do list and task management tool integrated into Gmail and Google Calendar. It helps you organize, track, and complete tasks across your Google ecosystem. - [Linear](https://composio.dev/toolkits/linear) - Linear is a modern issue tracking and project planning tool for fast-moving teams. It helps streamline workflows, organize projects, and boost productivity. - [Jira](https://composio.dev/toolkits/jira) - Jira is Atlassian’s platform for bug tracking, issue tracking, and agile project management. It helps teams organize work, prioritize tasks, and deliver projects efficiently. - [Clickup](https://composio.dev/toolkits/clickup) - ClickUp is an all-in-one productivity platform for managing tasks, docs, goals, and team collaboration. It streamlines project workflows so teams can work smarter and stay organized in one place. - [Monday](https://composio.dev/toolkits/monday) - Monday.com is a customizable work management platform for project planning and collaboration. It helps teams organize tasks, automate workflows, and track progress in real time. - [Addressfinder](https://composio.dev/toolkits/addressfinder) - Addressfinder is a data quality platform for verifying addresses, emails, and phone numbers. It helps you ensure accurate customer and contact data every time. - [Agiled](https://composio.dev/toolkits/agiled) - Agiled is an all-in-one business management platform for CRM, projects, and finance. It helps you streamline workflows, consolidate client data, and manage business processes in one place. - [AirOps MCP](https://composio.dev/toolkits/airops_mcp) - AirOps MCP is AirOps' managed content platform exposing brand context, knowledge bases, and workflows. Use it to give agents brand-aware content, automate content workflows, and track Brand Kit edits. - [Amazing Marvin](https://composio.dev/toolkits/amazing_marvin) - Amazing Marvin is a personal productivity app for tasks, projects, habits, planning, time tracking, and rewards. It helps you build a flexible workflow that matches how you actually work. - [AnyDB](https://composio.dev/toolkits/any_db) - AnyDB is a flexible collaborative database platform for organizing records, files, and automated workflows. It helps teams centralize data, collaborate on records, and automate routine tasks. - [ApptiveGrid](https://composio.dev/toolkits/apptivegrid) - ApptiveGrid is a collaborative no-code database and app platform for spaces, grids, forms, and structured records. It helps teams organize operational data, collect inputs, and build lightweight business apps without custom backend work. - [Ascora](https://composio.dev/toolkits/ascora) - Ascora is a cloud-based field service management platform for service businesses. It streamlines scheduling, invoicing, and customer operations in one place. - [Atlassian MCP](https://composio.dev/toolkits/atlassian_mcp) - Atlassian Rovo MCP connects agents to Jira, Jira Service Management, Confluence, Bitbucket, Compass, and Teamwork Graph workflows. It lets teams surface and act on Atlassian data while respecting existing user permissions. - [Basecamp](https://composio.dev/toolkits/basecamp) - Basecamp is a project management and team collaboration tool by 37signals. It helps teams organize tasks, share files, and communicate efficiently in one place. - [Beeminder](https://composio.dev/toolkits/beeminder) - Beeminder is an online goal-tracking platform that uses monetary pledges to keep you motivated. Stay accountable and hit your targets with real financial incentives. - [Boxhero](https://composio.dev/toolkits/boxhero) - Boxhero is a cloud-based inventory management platform for SMBs, offering real-time updates, barcode scanning, and team collaboration. It helps businesses streamline stock tracking and analytics for smarter inventory decisions. ## Frequently Asked Questions ### What are the differences in Tool Router MCP and Wispr Flow MCP? With a standalone Wispr Flow MCP server, the agents and LLMs can only access a fixed set of Wispr Flow tools tied to that server. However, with the Composio Tool Router, agents can dynamically load tools from Wispr Flow and many other apps based on the task at hand, all through a single MCP endpoint. ### Can I use Tool Router MCP with Atomic Agent? Yes, you can. Atomic Agent fully supports MCP integration. You get structured tool calling, message history handling, and model orchestration while Tool Router takes care of discovering and serving the right Wispr Flow tools. ### Can I manage the permissions and scopes for Wispr Flow while using Tool Router? Yes, absolutely. You can configure which Wispr Flow scopes and actions are allowed when connecting your account to Composio. You can also bring your own OAuth credentials or API configuration so you keep full control over what the agent can do. ### How safe is my data with Composio Tool Router? All sensitive data such as tokens, keys, and configuration is fully encrypted at rest and in transit. Composio is SOC 2 Type 2 compliant and follows strict security practices so your Wispr Flow data and credentials are handled as safely as possible. --- [See all toolkits](https://composio.dev/toolkits) · [Composio docs](https://docs.composio.dev/llms.txt)