A side-by-side look at the Firecrawl MCP Server and Outlook Assistant MCP servers: transport, auth, tools, quality signals, and the config to install either one in Claude, Cursor, Windsurf or VS Code.
Verdict: choose Firecrawl MCP Server for search & data extraction work as a local process; choose Outlook Assistant for communication as a local process. You can also run both side by side in one mcpServers block.
FIRECRAWL_API_KEY.firecrawl_scrape, firecrawl_map, firecrawl_search.auth, list-events, create-event.Rows where the two servers differ are marked.
| Spec | ||
|---|---|---|
| Category | Search & Data Extraction | Communication |
| Quality signal | 90/100 · Excellent | 68/100 · Great |
| Install command | Confirmed | Confirmed |
| Transport | Local subprocess (stdio) | Local subprocess (stdio) |
| Auth | API key required | Not declared |
| Pricing | Freemium | Not declared |
| Required env vars | FIRECRAWL_API_KEY | None listed |
| Verified owner | Verified | Community listing |
| GitHub stars | 7,610 | 39 |
| Directory activity | 10 views 0 installs 0 upvotes | 1 views 0 installs 0 upvotes |
| Tools | 25 | 22 |
| Full listing | View Firecrawl MCP Server | View Outlook Assistant |
firecrawl_scrape
Retrieve and extract content from one supplied URL through Firecrawl. Use this when the request identifies a page and needs its content or defined fields. It can return markdown, HTML, links, screenshots, branding data, a targeted answer, or JSON matching a supplied schema; JSON is useful when the requested result has defined fields, while markdown preserves readable page content.
This tool operates on a known page. For a set of pages use `firecrawl_crawl`, and to discover page URLs use `firecrawl_map` or `firecrawl_search`. Options include JavaScript render delay, cache age, main-content filtering, PII redaction, and lockdown cache-only retrieval. Browser actions may change the live page when interactive actions are enabled.
Firecrawl may reuse recently indexed content instead of refetching the page, and the reuse window varies by domain. Set `maxAge: 0` to force a live fetch, or a smaller `maxAge` to bound how stale reused content may be. A successful response does not by itself confirm that the state it describes is still current.
Returns the selected content formats and page metadata.
firecrawl_map
Enumerate URLs indexed under one website through Firecrawl without fetching each page's content. Use this when the request asks for a site's URL inventory, when several relevant pages must be located, or when the desired page URL is unknown. An optional `search` term narrows the URL list, while sitemap, subdomain, query-parameter, and result-limit options control coverage.
Returns matching URLs rather than page bodies. Retrieve one page with `firecrawl_scrape`; collect content across multiple pages with `firecrawl_crawl`.
firecrawl_search
Search web, news, or image sources and return ranked results. Operators include quoted phrases, `-term`, `site:host`, `inurl:term`, `intitle:term`, and `related:host`; the set is non-exhaustive. `includeDomains` and `excludeDomains` are mutually exclusive hostname filters; categories limit results to GitHub, research, PDF, or developer sources.
For a programming question, add `categories: ["developer"]`. It searches an index of GitHub issues, merged pull requests, repository READMEs, and curated documentation sites, and returns the hits in `data.developer` beside the web results.
`categories: ["research"]` restricts these web results to research-affiliated websites and returns page snippets. The `firecrawl_research_*` tools are a separate surface that searches paper abstracts and full text across biomedical (PubMed, bioRxiv, medRxiv) and arXiv literature.
`scrapeOptions` can attach extracted page content; pages fetched this way use a fixed reuse window and ignore `maxAge`, so use `firecrawl_scrape` when a live fetch is required. Returns source-type result groups and usage metadata. Authenticated responses can include an `id` for optional search feedback.
firecrawl_crawl
Start a multi-page crawl at a website URL, poll it to a terminal state, and return the final status and collected data. Scope can be bounded with include/exclude paths, depth, page limit, subdomain/external-link controls, sitemap handling, delay, and scrape options.
Crawl results can be large; use conservative limits when full-site coverage is unnecessary. Webhooks and interactive scrape actions are unavailable in safe mode. Returns the crawl ID, status, and page data.
firecrawl_check_crawl_status
Retrieve the current status, progress, and available results for an existing crawl ID. This only reads Firecrawl job state and does not start or modify the crawl.
firecrawl_extract
Deprecated compatibility entry point. Use firecrawl_scrape once per known URL with formats: ["json"] and jsonOptions containing the prompt and schema. Use firecrawl_search or firecrawl_agent before Scrape when URLs are not known.
firecrawl_agent
Start an asynchronous web research job from a prompt, optional seed URLs, and an optional JSON schema. Use this for a requested synthesis across multiple sources when the task can wait for asynchronous completion. The agent can search, navigate, read pages, and assemble a structured result.
This call returns only a job ID, not the research result. Read the job with `firecrawl_agent_status` until it reaches `completed` or `failed`; research commonly takes several minutes. If the job cannot finish within the task's available time, `firecrawl_search` and `firecrawl_scrape` can gather evidence synchronously.
firecrawl_agent_status
Retrieve progress or final results for a `firecrawl_agent` job ID. A `processing` response is non-terminal and does not contain the final research result. Check again after 15–30 seconds until the status is `completed` or `failed`; complex jobs can take several minutes. If the job cannot finish within the task's available time, use `firecrawl_search` and `firecrawl_scrape` to complete the requested output.
Returns job status, progress information, and result data when completed.
firecrawl_interact
Open or reuse a live browser session to navigate a page, click controls, fill fields, or run browser code. Provide either `url` or `scrapeId`, and either a natural-language `prompt` or executable `code`; code can run as Bash, Python, or Node with a bounded timeout.
This acts on the live site, so actions such as form submission can create persistent external side effects. Returns execution output, stdout/stderr, exit status, and session viewing URLs.
firecrawl_interact_stop
Stop the live interact session associated with a `scrapeId` and release its resources. Returns a success confirmation.
firecrawl_parse
Parse one supported document into markdown, HTML, links, summary, targeted answers, or JSON matching a schema. Supported inputs include common HTML, PDF, Word, RTF, OpenDocument, and spreadsheet files; PDF parsing can be bounded with `pdfOptions.maxPages`.
Local MCP reads `filePath` from the server filesystem. Hosted MCP uses two calls: first provide `filePath` to receive upload instructions, upload locally, then call again with the returned `uploadRef`; do not send both fields together. Remote web URLs belong in `firecrawl_scrape`.
Set `redactPII` to request redaction of personally identifiable information in the returned content. `zeroDataRetention` requires an eligible authenticated account; omit it for anonymous keyless use. Returns upload instructions for hosted phase one or parsed document content for the final call.
firecrawl_monitor_create
Create a recurring scrape, crawl, or search monitor that compares each check with its retained predecessor. The simple form accepts `page`/`pages` or `queries` plus a plain-language `goal`; the advanced `body` form controls targets, schedule, change-tracking formats, judging, retention, webhook, and notifications.
In the simple form, a `goal` is required. If `queries` contains one or more non-empty values and is supplied with `page`/`pages`, `queries` create the search target and page targets are ignored. A monitor schedules future network checks and can send configured email or webhook notifications. Returns the created monitor.
authManage authentication with the Microsoft Graph API. action=`status` (default) returns the current auth state, refreshing an expired access token while the refresh token is still valid (~90 days); call it first to check before other tools. action=`authenticate` starts sign-in: `method: "device-code"` (default, works headlessly) returns a code + URL for the user to visit; `method: "browser"` uses the local auth server on :3333 (run `npm run auth-server` first). `force: true` re-authenticates over an existing valid session. If sign-in reports that OUTLOOK_CLIENT_ID is not configured, ask the user for their Azure Application (client) ID and pass it as `clientId`. action=`device-code-complete` finishes device-code sign-in once the browser shows it succeeded. action=`about` returns server version, configured audience, scope list and other diagnostics. Tokens persist to `~/.outlook-assistant-tokens.json` and survive server restarts.list-eventsList calendar events for the signed-in user (read-only). By default returns upcoming events (start ≥ now). Optional `startAfter`, `startBefore` and `subject` filters find past, current or specifically-named events; supplying any of them replaces the default "now" lower bound and the filters are AND-ed together. Results are oldest first, except when the search only looks backwards (`startBefore` without `startAfter`, or `subject` alone), where they are newest first. Each event shows its subject, location, start/end, a body preview and its id. Use `count` (default 10, max 100) to control page size. Each start/end is returned as a canonical UTC ISO-8601 instant (e.g. `2026-04-02T22:00:00.000Z`) followed by a labelled local rendering in the configured display timezone (default Australia/Melbourne; override with `OUTLOOK_DEFAULT_TIMEZONE`) — the UTC value is authoritative, so consumers never have to guess the zone.create-eventCreate a new calendar event on the signed-in user's default calendar. Returns the created event with its `id`, `webLink`, and (if attendees are present) an auto-generated online-meeting URL — attendees receive invitations on save, so pass `dryRun: true` first to preview who would be invited (and how many are external) without creating anything. When `OUTLOOK_ALLOWED_RECIPIENTS` is set, every attendee must be allowed or nothing is created. Times use the configured timezone (default Australia/Melbourne; override with `OUTLOOK_DEFAULT_TIMEZONE`); omit the `Z` suffix to send local time. Use `manage-event` action=`update` to modify an event after creation, or `manage-event` action=`cancel`/`delete` to remove it. Real creates count against the session limit (`OUTLOOK_MAX_CREATE_EVENT_PER_SESSION`, else `OUTLOOK_MAX_EMAILS_PER_SESSION`); 0 refuses every create.Paste either block (or both) into your client config, e.g. claude_desktop_config.json or ~/.cursor/mcp.json. We only show a config when the install command is confirmed.
{
"mcpServers": {
"firecrawl-firecrawl-mcp-server": {
"command": "npx",
"args": [
"-y",
"firecrawl-mcp"
],
"env": {
"FIRECRAWL_API_KEY": "YOUR_FIRECRAWL_API_KEY_HERE"
}
}
}
}
{
"mcpServers": {
"littlebearapps-outlook-assistant": {
"command": "npx",
"args": [
"-y",
"@littlebearapps/outlook-assistant"
]
}
}
}
manage-eventManage an existing calendar event. `dryRun: true` previews any action without changing or sending anything: who would be emailed, with an external count (update also: who is added or removed, and the PATCH body). action=`update` edits fields via PATCH (subject, start, end, attendees, body, location, isOnlineMeeting, sensitivity, showAs, importance, categories, reminderMinutesBeforeStart); only fields you pass change. action=`decline` declines an invitation (optional `comment`; `sendResponse: false` declines without notifying the organiser). action=`cancel` cancels an event you organised and emails attendees. action=`delete` removes the event from your calendar (Graph documents no guaranteed recovery); deleting a meeting you organised that has attendees still emails them a cancellation, so use `cancel` with a `comment` to control that message. Returns the updated event on update; a confirmation otherwise. There is no `accept` action: accept invitations in the Outlook UI, as Graph's accept verb is unreliable.search-emailsSearch, list, delta-sync or thread-group emails (read-only); parameters set the mode. No params: recent emails in `folder` (default `inbox`). `query`/`from`/`to`/`subject`/date filters: search, combined as an OData filter. `searchExpression` (deprecated alias `kqlQuery`): a raw Graph `$search` expression. `deltaMode: true`: current state plus a `deltaToken` to pass back next time for changes only. `groupByConversation: true`: conversation threads. `conversationId`: every message in one thread. `internetMessageId`: the message with that RFC Message-ID. `sharedMailbox` (alias `email`) searches a shared/delegated mailbox, custom folders and nested paths included. Personal Outlook.com accounts have limited `$search`, so the tool falls back to OData filters and a recent listing automatically; structured filters (`from`/`subject`/`receivedAfter`/`hasAttachments`/`unreadOnly`) give cleaner results there. Returns up to `count` messages (id/subject/from/receivedDateTime/preview); `outputVerbosity` expands them.read-emailRead a single email by id (read-only). Returns subject, from/to/cc, date and the body as Markdown (HTML stripped to text): up to 2,000 characters by default, up to 40,000 with `outputVerbosity: full`; a cut body ends with a note on how to get the rest. With `headersMode: true`: returns RFC-822 forensic headers in place of the body (DKIM, SPF, DMARC, Received chain, Message-ID, Authentication-Results) — `importantOnly: true` for the security-relevant subset, `groupByType: true` for a category-bucketed view, `raw: true` for JSON. With `includeHeaders: true` (non-headers-mode): adds basic headers alongside the body. **If the id came from a shared/delegated mailbox (e.g. via `search-emails` or `access-shared-mailbox` with `sharedMailbox` set), you MUST pass the same `sharedMailbox` (or alias `email`) here** — message IDs are mailbox-scoped, and reading a shared-mailbox id without it fails with 404 ErrorInvalidMailboxItemId.send-emailCompose and send an email immediately (destructive: sends external comms). Returns a confirmation. Safety controls: `dryRun: true` returns the composed message for review without sending; `checkRecipients: true` runs `get-mail-tips` first and returns its warnings. If the tips show an out-of-office reply, a full mailbox, a delivery restriction or external recipients, the send is refused until repeated with `acknowledgeWarnings: true`. Personal Outlook.com accounts return no tips, and no warnings is not proof of delivery. Subject to the session limit (`OUTLOOK_MAX_SEND_EMAIL_PER_SESSION`, else `OUTLOOK_MAX_EMAILS_PER_SESSION`; 0 refuses every send) and recipient allowlist (`OUTLOOK_ALLOWED_RECIPIENTS`) when configured; both refuse before any Graph request. For a review-before-send workflow, use `draft` (action=`create` → `update` → `send`); a draft can be checked in Outlook before it goes. Comma-separated recipient strings or arrays both accepted.draftDraft lifecycle for review-before-send workflows (destructive: covers `send` and `delete`). action=`create` saves a new draft and returns its id (`dryRun: true` previews without saving; `checkRecipients: true` runs mail-tips first). action=`update` patches a draft by `id` (only fields passed change). action=`send` sends a draft and counts against the `send-email` session limit (0 refuses every send). action=`delete` moves a draft to Recoverable Items (restorable for a limited time). update/send/delete refuse any `id` that is not an unsent draft. action=`reply`/`reply-all` creates a reply draft from a message `id` (`comment` prepends text; not with `body`). action=`forward` creates a forward draft (needs `id` and `to`). The recipient allowlist (OUTLOOK_ALLOWED_RECIPIENTS) applies to create/update/forward, to the draft's current to/cc/bcc on send, and to reply/reply-all, whose draft is deleted if a recipient is not allowed. Returns the draft on create/update/reply/forward; a status on send/delete.update-emailUpdate message state without modifying content (idempotent — safe to retry). action=`mark-read`/`mark-unread` sets `isRead` on a single message by `id`. action=`flag` sets a follow-up flag with optional `dueDateTime`/`startDateTime`: ISO 8601 with a time, kept as that exact instant when it has `Z` or a ±hh:mm offset and read in OUTLOOK_DEFAULT_TIMEZONE when it has none; date-only or unparseable values are refused before any change. With only `dueDateTime`, the start is 09:00 on the due date, or the due time if earlier. action=`unflag` clears the flag; action=`complete` marks it done. Flag/unflag/complete take `id` (single) or `ids` (batch, updated one at a time: one PATCH each, not Graph `$batch`). `sharedMailbox` (alias `email`) updates messages in a shared/delegated mailbox (default: the signed-in account; needs Mail.ReadWrite.Shared and delegate access). Returns a status per message.attachmentsInspect or retrieve email attachments. action=`list` (default) returns metadata for all attachments on `messageId` (id, name, contentType, size, isInline) — read-only. action=`view` returns inline content for text/JSON/XML attachments via `attachmentId`; binary types require download. action=`download` saves the attachment under a new, unique name in `outputDir` (default system temp directory, auto-created; must be inside the temp directory, ~/Downloads, ~/Documents or OUTLOOK_EXPORT_DIR) and returns the saved file path. `messageId` is required for all actions; `attachmentId` is required for view/download. If `messageId` came from a shared/delegated mailbox, pass the same `sharedMailbox` (or alias `email`) — attachment IDs are scoped to the message and fail under /me otherwise. Use `outputVerbosity` to control list field count.exportExport emails to files. target=`message` (default) exports one email by `id` (mime/eml/markdown/json/csv) to `savePath`: a directory (or a path ending in `/`, created if missing) gets a new, unique file name; a file path is created new and an existing file is replaced only with `overwrite: true`. target=`messages` batch-exports `emailIds`, or matches for `searchQuery`/`query`, into `outputDir`, at most 100 messages per call. target=`conversation` exports a thread (up to 1000 messages) by `conversationId` into `outputDir` (eml/mbox/markdown/json/html/csv; `order: "reverse"` for newest first). target=`mime` returns raw RFC-822 MIME for `id` (`headersOnly`, `base64`, `maxSize`, default 1MB). Files are written only inside the system temp directory (the default), ~/Downloads, ~/Documents or OUTLOOK_EXPORT_DIR, never to dot-prefixed names. Pass `sharedMailbox` (alias `email`) when the ids come from a shared mailbox. `includeAttachments` defaults to true for one message, false for batch.get-mail-tipsPre-send recipient validation via Graph `POST /me/getMailTips` (read-only; uses the existing `Mail.Read` scope — no extra permissions). Returns per-recipient tips covering automatic replies (out-of-office), mailbox full status, custom admin mail tips, delivery restrictions, moderation requirements, external-vs-internal scope, max message size, and group member counts (total + external). Use ahead of `send-email` or `draft` action=`create` to catch issues like OOO replies or external-recipient warnings before the message goes out; `send-email`/`draft` accept `checkRecipients: true` to invoke this automatically. Accepts either a comma-separated string or an array of addresses; `tipTypes` filters which tips are requested (defaults to all).