Community Gmail server with auto-auth: search, read and send email via MCP.
Nothing to run from here
Proton Mail is installed from its publisher's own source and answers where it runs, so this marketplace is not in the path of a single call. There is no address here to send one to, and a panel that pretended otherwise would be showing you an answer we made up. Install it and call it from your own client — the Installation tab has the entry for each one.
96/100
trust score
9
downloads a day
local
runs on your machine
391
downloads in 30d
What it does
Unofficial MCP server for Proton Mail (not affiliated with Proton AG) — send, read, search, and organize email over SMTP and IMAP
Quickstart
# 1 — add it to Claude Code
claude mcp add --scope user --env 'PROTONMAIL_USERNAME=${PROTONMAIL_USERNAME}' --env 'PROTONMAIL_PASSWORD=${PROTONMAIL_PASSWORD}' --transport stdio proton-mail-mcp '--' npx -y proton-mail-mcp
# 2 — ask your agent for something these tools do
# send_email, reply_email, reply_all_email and 29 more
It needs PROTONMAIL_USERNAME, PROTONMAIL_PASSWORD: this reads them from your environment, so export them first.
Proton Mail costs nothing on mcprush: there is no plan to choose here, no cap to set and nothing mcprush can bill you for.
Publishernot claimed
SEAll 2 servers
sethbang2 servers
2 categories
Collected from a public index. Nobody has claimed this account, so nothing here was written by its author — claim it if it is yours.
Installation
Where are you running it?
Every route below installs the same thing and ends at the same approval screen. This one runs on your machine: your client starts Proton Mail as a process under your own user, with your files and your network, so the tool surface below is what it can do to your computer rather than to a server somewhere else. Its scan result is on the Security tab — read the surface before you approve it.
Claude Codedeclared by this listing
This is a public server: you run it yourself and this marketplace is not in the path. You need Node.js — the current LTS (24) is the safe choice; npx comes with it. Claude Code registers it in one command; --scope user makes it available in every project. Claude Code reads PROTONMAIL_USERNAME (Your Proton Mail email address), PROTONMAIL_PASSWORD (Your Proton Mail SMTP password (not your login password; generate one in Proton settings)) from your environment each time it starts the server, so the value stays out of its config: export them in your shell profile (macOS, Linux), or run setx PROTONMAIL_USERNAME …, setx PROTONMAIL_PASSWORD … and open a new terminal (Windows). The '--' is in quotes so that PowerShell passes it on.
claude mcp add --scope user --env 'PROTONMAIL_USERNAME=${PROTONMAIL_USERNAME}' --env 'PROTONMAIL_PASSWORD=${PROTONMAIL_PASSWORD}' --transport stdio proton-mail-mcp '--' npx -y proton-mail-mcpReconnect, or start a new session, and the tools appear in the model’s tool list.
Free on mcprush, and it runs from its publisher’s own package: this marketplace is not in the path, so there is no key of ours to hold and no card to add.
What it asks of you
It reaches a system of yours, so it needs your own credential rather than the publisher’s: PROTONMAIL_USERNAME — Your Proton Mail email address; PROTONMAIL_PASSWORD — Your Proton Mail SMTP password (not your login password; generate one in Proton settings). You set them in your own client’s config, on your machine — this marketplace never holds it.
What it writes
One config entry your client uses to start the process locally. A local server runs with your file system and your network, which is why it is priced without metering.
What you approve
32 tools, with what each one reads, writes and reaches shown before you agree — the same list on every route above. Read the tool surface.
Tool surface
What the model actually sees. Descriptions are diffed on every release — see version history.
Send an email using Proton Mail SMTP. HTML bodies are sanitized through a conservative allowlist by default (v1.0.0: sanitizeHtml defaults to true) — scripts, event handlers, inline styles, and remote <img> beacons are stripped. Pass sanitizeHtml: false to send full-fidelity HTML in…
No parameter schema on file.
Reply to an email message. Reads the original message and sends a reply with proper threading headers (In-Reply-To, References). Response leads with a [sent-copy:verified|unverified] token; the [reply-to:*] family of tokens does NOT apply here because this tool doesn't accept a replyTo…
No parameter schema on file.
Reply to all recipients of an email (sender + original TO + original CC), excluding the authenticated user. Sends with proper threading headers. Equivalent to reply_email with replyAll: true, exposed as a dedicated tool for discoverability. Response leads with…
No parameter schema on file.
Forward an email message. Reads the original message and sends it to new recipients with proper threading headers. Response leads with [sent-copy:verified|unverified]; the [reply-to:*] tokens do not apply because this tool has no replyTo parameter to verify. sanitizeHtml scope: the…
No parameter schema on file.
List available email folders/mailboxes with message counts. The per-folder counts come from a cached IMAP STATUS that Proton Mail Bridge can serve stale — do NOT treat them as authoritative for decisions like "is this folder empty before deleting". Use count_messages or folder_stats (both…
No parameter schema on file.
List recent messages from an email folder, sorted by date (newest first). Returns subject, sender, date, and flags for each message. A non-selectable namespace container (e.g. Folders/Labels) is rejected with an actionable error rather than returning an empty list. Pagination note: the…
No parameter schema on file.
Read a specific email message by UID. Returns headers and body content. By default prefers the plain-text part and strips HTML tags from HTML-only messages. Body is truncated to avoid exceeding token limits (default 50 000 chars). ⚠️ Prompt-injection caveat (agentic readers). The returned…
No parameter schema on file.
List attachment metadata for a message without downloading the body. Returns part numbers, filenames, content types, and sizes — use these with download_attachment to fetch the content.
No parameter schema on file.
Download an email attachment by part number. Use read_message or list_attachments first to see available attachments and their part numbers. By default returns base64-encoded content inline (read-only). When saveTo is provided AND the ALLOW_FILE_DOWNLOAD_DIR env var is set, this tool WRITES the…
No parameter schema on file.
Search for messages in a folder by various criteria (sender, subject, date, flags). Returns matching message summaries sorted by date (newest first). Note: recently sent or received messages may take a few seconds to become searchable by subject or body due to server-side indexing delays;…
No parameter schema on file.
Move an email message to a different folder. Note: the message gets a new UID in the destination folder — the original UID is no longer valid after the move. UID + folder pair caveat: IMAP UIDs are per-folder, so UID 42 in INBOX and UID 42 in Sent identify different messages. Always carry the…
No parameter schema on file.
Delete an email message. By default moves to Trash for safety; set permanent=true to permanently expunge. Note: moving to Trash assigns a new UID in the Trash folder — the original UID is no longer valid. UID + folder pair caveat: IMAP UIDs are per-folder. Always pair a UID with the folder it…
No parameter schema on file.
Add or remove flags on an email message. System flags (RFC 3501): \\Seen (read), \\Flagged (starred), \\Answered, \\Draft, \\Deleted, \\Recent. User-defined keywords without a backslash prefix are also accepted (alphanumeric + underscore, e.g. "Important", "Custom_Tag"), but Proton Mail Bridge has…
No parameter schema on file.
Get all messages in a conversation thread by walking In-Reply-To and References headers. Returns messages sorted chronologically (oldest first).
No parameter schema on file.
Mark all unread messages in a folder as read. Optionally limit to messages older than a given date. Pass dryRun: true to preview the affected count without flipping any flags.
No parameter schema on file.
Save an email as a draft without sending it. The draft is placed in the user's \Drafts special-use folder (resolved at runtime; falls back to literal Drafts if no annotation). The destination is intentionally not caller-controlled — prior versions accepted an arbitrary folder parameter that…
No parameter schema on file.
Move multiple messages to a different folder in one operation. Provide EITHER uids (an explicit list) OR match (search criteria — same shape as search_messages), not both. Set dryRun: true to preview what would be moved without making changes. Note: moved messages get new UIDs in the…
No parameter schema on file.
Delete multiple messages in one operation. Provide EITHER uids OR match. By default soft-deletes to Trash; pass permanent: true to expunge. permanent: true ALSO requires confirm: true (the expunge is irreversible — there is no Trash to recover from). dryRun: true previews without…
No parameter schema on file.
Add or remove flags on multiple messages in one operation. Provide EITHER uids OR match, plus at least one of flagsToAdd / flagsToRemove. Same flag whitelist as update_message_flags.
No parameter schema on file.
Create a new mailbox folder. Returns gracefully if the folder already exists. On Proton Mail, folders must be created under the "Folders/" namespace (e.g. "Folders/Receipts") — root-level paths are rejected by the server with an actionable error.
No parameter schema on file.
Create a new Proton label. Pass the bare label name (e.g. "Important") — the tool prepends the "Labels/" namespace internally. Labels are non-exclusive tags: a message can carry many labels in addition to living in one folder. Apply or remove labels on messages with update_message_labels.…
No parameter schema on file.
Rename a mailbox folder or label. Errors if the source path does not exist. Works for both "Folders/" and "Labels/" paths.
No parameter schema on file.
Delete a mailbox folder or label container. Restricted to the "Folders/" and "Labels/" namespaces to protect system mailboxes (INBOX, Sent, Trash, etc.). On Proton Mail this is not a destructive message operation: deleting a folder relocates its contents into "All Mail"; deleting a label…
No parameter schema on file.
Add or remove Proton labels on a message. Labels live under the "Labels/" namespace and are additive — the message stays in its source folder while gaining or losing label tags. Pass full paths in labelsToAdd / labelsToRemove (e.g. ["Labels/Important", "Labels/Work"]). Adds are strict:…
No parameter schema on file.
Add or remove Proton labels on many messages in one operation. Provide EITHER uids OR match (XOR), plus at least one of labelsToAdd / labelsToRemove. Same label-path rules as update_message_labels (must start with "Labels/"). Supports dryRun: true for safe preview.
No parameter schema on file.
Permanently delete ALL messages in a folder (atomic UID EXPUNGE via messageDelete). Requires confirm: true (skipped for dryRun). By default restricted to Trash/Junk; pass allowAnyFolder: true to empty other folders. THIS IS NOT REVERSIBLE. Pass dryRun: true to preview the count that would…
No parameter schema on file.
Count messages in a folder matching optional search criteria. Returns just a number (no envelopes fetched). The attachment filters (hasAttachment, attachmentName, attachmentType) are rejected here — they require an envelope scan that defeats the count's speed promise. Use search_messages for…
No parameter schema on file.
Return aggregate stats for a folder: total/unread (free), plus scanned-envelope aggregations (oldest/newest/total bytes). Default scanLimit 5000, max 20000. Response always includes scanned/truncated so callers can detect partial results. A non-selectable namespace container (e.g.…
No parameter schema on file.
Return a frequency table of top senders for a folder, optionally filtered by date range. Buckets are keyed by lowercased email address. Default limit 20, scanLimit 5000 (max 20000). Each row carries a direction of "self" or "received" so callers can distinguish messages from the authenticated…
No parameter schema on file.
Move every message in a thread to a destination folder. By default acts only in the seed message's folder; pass acrossFolders:true to walk INBOX/Sent/All Mail. dryRun:true previews the affected per-folder UIDs without moving.
No parameter schema on file.
Delete every message in a thread. Default soft-deletes to Trash; permanent:true expunges. acrossFolders:false by default for safety. dryRun:true previews.
No parameter schema on file.
Add or remove flags on every message in a thread, identified by Message-ID. Use this instead of update_message_flags when you want the change applied to a whole conversation, or bulk_update_flags when you have a flat set of UIDs rather than a thread. At least one of flagsToAdd/flagsToRemove must…
No parameter schema on file.
Free
$0 forever
- Every tool, no call limit
- No card, and mcprush charges nothing for it
- Source published under MIT
- Runs on your machine — nothing of it reaches our gateway
- Nothing to cap, because mcprush bills nothing
What counts against your monthly calls
| Tool | Unit | Calls used | Out of the allowance |
|---|
Nothing here is billable: mcprush charges nothing to install Proton Mail or to call it, at any volume.
Call allowance · this installenforced at the gateway
Monthly allowance, in calls
—
—
Alert at 80% · then the tool returns a normal MCP error
Used this month: —
ATrust score 96 / 100
Release 1.0.2 · scanned 15 Oct 2026
Two independent axes, because powerful and malicious are different questions. The grade is threat only. The capability level is blast radius, and it is never a penalty on the grade — it is priced as one subtract-only term in the score, where you can see it.
Threat score99
Client score96
Findings1 threat · 1 capability
Capabilitymoderate
Coveragesource
Verificationsource
Surface digeste857f371ca7cf4db7c672a5cec083c37f997010868a6a7d6bbff75db58967322
Score vectorthreat 99 → client 96
| Term | Level | What it prices | Points |
|---|---|---|---|
| capability-exposure | moderate | capability blast radius (moderate) — client exposure if the model is manipulated | −3 |
| verification-discount | source | publisher verification (provenance) — cryptographic build provenance ties the artifact to its source | −0 |
| coverage-honesty | source | inspection depth (source) — how much of the target the scan could see | −0 |
score = clamp(round(threatScore − E_cap − E_ver − E_cov), 0, 100). No model, no opinion, no network call at scoring time: the same methodology version against the same bytes always produces this number, and every line above is a field of the report rather than a summary of it.
What the scan could actually read
A grade is only as meaningful as its coverage, so the scanner publishes its own depth before it publishes its result.
Tool surfaceread
Implementation sourceread
Package metadataread
Live transportnot available
Tools were statically extracted from the published source (32 recovered), not enumerated from a running server. Tool-poisoning, Unicode-smuggling, capability and toxic-flow analysis ran on this inferred surface, but a mis-parsed registration could be missed or mis-attributed, so tool-derived findings are capped below “confirmed”. To grade the real runtime surface, scan the running server: --command "npx -y <package>".
1 of the 32 recovered tools are registered only when !READONLY holds (build/index.js:496) — typically when a key or option is configured. They are part of the server's surface and were analysed; a server started without that configuration exposes only the other 10.
1 of the 32 recovered tools are registered only when !READONLY holds (build/index.js:692) — typically when a key or option is configured. They are part of the server's surface and were analysed; a server started without that configuration exposes only the other 10.
1 of the 32 recovered tools are registered only when !READONLY holds (build/index.js:842) — typically when a key or option is configured. They are part of the server's surface and were analysed; a server started without that configuration exposes only the other 10.
1 of the 32 recovered tools are registered only when !READONLY holds (build/index.js:966) — typically when a key or option is configured. They are part of the server's surface and were analysed; a server started without that configuration exposes only the other 10.
1 of the 32 recovered tools are registered only when !READONLY holds (build/index.js:1589) — typically when a key or option is configured. They are part of the server's surface and were analysed; a server started without that configuration exposes only the other 10.
1 of the 32 recovered tools are registered only when !READONLY holds (build/index.js:1628) — typically when a key or option is configured. They are part of the server's surface and were analysed; a server started without that configuration exposes only the other 10.
1 of the 32 recovered tools are registered only when !READONLY holds (build/index.js:1709) — typically when a key or option is configured. They are part of the server's surface and were analysed; a server started without that configuration exposes only the other 10.
1 of the 32 recovered tools are registered only when !READONLY holds (build/index.js:1861) — typically when a key or option is configured. They are part of the server's surface and were analysed; a server started without that configuration exposes only the other 10.
1 of the 32 recovered tools are registered only when !READONLY holds (build/index.js:1901) — typically when a key or option is configured. They are part of the server's surface and were analysed; a server started without that configuration exposes only the other 10.
1 of the 32 recovered tools are registered only when !READONLY holds (build/index.js:2054) — typically when a key or option is configured. They are part of the server's surface and were analysed; a server started without that configuration exposes only the other 10.
1 of the 32 recovered tools are registered only when !READONLY holds (build/index.js:2111) — typically when a key or option is configured. They are part of the server's surface and were analysed; a server started without that configuration exposes only the other 10.
1 of the 32 recovered tools are registered only when !READONLY holds (build/index.js:2195) — typically when a key or option is configured. They are part of the server's surface and were analysed; a server started without that configuration exposes only the other 10.
1 of the 32 recovered tools are registered only when !READONLY holds (build/index.js:2284) — typically when a key or option is configured. They are part of the server's surface and were analysed; a server started without that configuration exposes only the other 10.
1 of the 32 recovered tools are registered only when !READONLY holds (build/index.js:2309) — typically when a key or option is configured. They are part of the server's surface and were analysed; a server started without that configuration exposes only the other 10.
1 of the 32 recovered tools are registered only when !READONLY holds (build/index.js:2341) — typically when a key or option is configured. They are part of the server's surface and were analysed; a server started without that configuration exposes only the other 10.
1 of the 32 recovered tools are registered only when !READONLY holds (build/index.js:2363) — typically when a key or option is configured. They are part of the server's surface and were analysed; a server started without that configuration exposes only the other 10.
1 of the 32 recovered tools are registered only when !READONLY holds (build/index.js:2398) — typically when a key or option is configured. They are part of the server's surface and were analysed; a server started without that configuration exposes only the other 10.
1 of the 32 recovered tools are registered only when !READONLY holds (build/index.js:2456) — typically when a key or option is configured. They are part of the server's surface and were analysed; a server started without that configuration exposes only the other 10.
1 of the 32 recovered tools are registered only when !READONLY && ALLOW_EMPTY_FOLDER holds (build/index.js:2548) — typically when a key or option is configured. They are part of the server's surface and were analysed; a server started without that configuration exposes only the other 10.
1 of the 32 recovered tools are registered only when !READONLY holds (build/index.js:2769) — typically when a key or option is configured. They are part of the server's surface and were analysed; a server started without that configuration exposes only the other 10.
1 of the 32 recovered tools are registered only when !READONLY holds (build/index.js:2830) — typically when a key or option is configured. They are part of the server's surface and were analysed; a server started without that configuration exposes only the other 10.
1 of the 32 recovered tools are registered only when !READONLY holds (build/index.js:2896) — typically when a key or option is configured. They are part of the server's surface and were analysed; a server started without that configuration exposes only the other 10.
Capability — what it could do if the model were manipulated
Tags derived from each tool’s schema and the implementation, not from what the tool calls itself. moderate is the level these add up to.
10 tools are registered unconditionally · (1) 1 more only when !READONLY holds (build/index.js:496) · (2) 1 more only when !READONLY holds (build/index.js:692) · (3) 1 more only when !READONLY holds (build/index.js:842) · (4) 1 more only when !READONLY holds (build/index.js:966) · (5) 1 more only when !READONLY holds (build/index.js:1589) · (6) 1 more only when !READONLY holds (build/index.js:1628) · (7) 1 more only when !READONLY holds (build/index.js:1709) · (8) 1 more only when !READONLY holds (build/index.js:1861) · (9) 1 more only when !READONLY holds (build/index.js:1901) · (10) 1 more only when !READONLY holds (build/index.js:2054) · (11) 1 more only when !READONLY holds (build/index.js:2111) · (12) 1 more only when !READONLY holds (build/index.js:2195) · (13) 1 more only when !READONLY holds (build/index.js:2284) · (14) 1 more only when !READONLY holds (build/index.js:2309) · (15) 1 more only when !READONLY holds (build/index.js:2341) · (16) 1 more only when !READONLY holds (build/index.js:2363) · (17) 1 more only when !READONLY holds (build/index.js:2398) · (18) 1 more only when !READONLY holds (build/index.js:2456) · (19) 1 more only when !READONLY && ALLOW_EMPTY_FOLDER holds (build/index.js:2548) · (20) 1 more only when !READONLY holds (build/index.js:2769) · (21) 1 more only when !READONLY holds (build/index.js:2830) · (22) 1 more only when !READONLY holds (build/index.js:2896) — typically when a key or option is configured. All 32 are in the table below and were analysed; the conditional ones are marked.
| Tool | Capability tags | Why the tag was assigned |
|---|---|---|
| send_email conditional (1) | external-sinkstate-write | external-sink: keyword "send_email"; state-write: declares readOnlyHint: false |
| reply_email conditional (2) | state-write | state-write: declares readOnlyHint: false |
| reply_all_email conditional (3) | state-write | state-write: declares readOnlyHint: false |
| forward_email conditional (4) | external-sinkstate-write | external-sink: keyword "forward_email"; state-write: declares readOnlyHint: false |
| list_folders | no tags | |
| list_messages | no tags | |
| read_message | untrusted-input | untrusted-input: keyword "read_message" |
| list_attachments | untrusted-input | untrusted-input: keyword "fetch" |
| download_attachment | untrusted-inputstate-write | untrusted-input: keyword "download"; state-write: declares readOnlyHint: false |
| search_messages | no tags | |
| move_message conditional (5) | state-write | state-write: declares readOnlyHint: false, destructiveHint: true |
| delete_message conditional (6) | state-write | state-write: declares readOnlyHint: false, destructiveHint: true |
| update_message_flags conditional (7) | state-write | state-write: declares readOnlyHint: false |
| get_thread | no tags | |
| mark_all_read conditional (8) | untrusted-inputstate-write | untrusted-input: keyword "read_message"; state-write: declares readOnlyHint: false |
| save_draft conditional (9) | state-write | state-write: declares readOnlyHint: false |
| bulk_move conditional (10) | state-write | state-write: declares readOnlyHint: false, destructiveHint: true |
| bulk_delete conditional (11) | state-write | state-write: declares readOnlyHint: false, destructiveHint: true |
| bulk_update_flags conditional (12) | state-write | state-write: declares readOnlyHint: false |
| create_folder conditional (13) | state-write | state-write: declares readOnlyHint: false |
| create_label conditional (14) | state-write | state-write: declares readOnlyHint: false |
| rename_folder conditional (15) | state-write | state-write: declares readOnlyHint: false, destructiveHint: true |
| delete_folder conditional (16) | state-write | state-write: declares readOnlyHint: false |
| update_message_labels conditional (17) | state-write | state-write: declares readOnlyHint: false |
| bulk_update_labels conditional (18) | state-write | state-write: declares readOnlyHint: false |
| empty_folder conditional (19) | state-write | state-write: declares readOnlyHint: false, destructiveHint: true |
| count_messages | no tags | |
| folder_stats | no tags | |
| top_senders | no tags | |
| move_thread conditional (20) | state-write | state-write: declares readOnlyHint: false, destructiveHint: true |
| delete_thread conditional (21) | state-write | state-write: declares readOnlyHint: false, destructiveHint: true |
| flag_thread conditional (22) | state-write | state-write: declares readOnlyHint: false |
Toxic-flow graph
The lethal trifecta, checked as a graph rather than as a checklist: untrusted input, a sensitive source and an external sink have to meet before there is a path worth worrying about.
flow-1
Untrusted input can drive an external action even though no sensitive source is exposed.
Supply chain and provenance
Integrityfirst-seen
Surface digeste857f371ca7cf4db
Scanned2026-10-15
This is the first scan of this surface here, so there is nothing yet to compare it against.
Every result on this tab comes from one deterministic pass over the published package — offline, rule by rule, and auditable line by line above. Same methodology version, same bytes, same score.
Release history
Always latest — the install command names no version, so it installs the newest release when it runs; the current release on file is 1.0.2. Pin a release below and the command asks for exactly that one.
1.0.2
undated
latestNo release note was published with this version.
—
no reviews yet
Who can review
One review per account, from a signed-in account that does not publish this listing. Nothing else is asked — you can say what you think before you install it. Publishers can reply once per review, and a review can be edited or deleted by whoever wrote it.
Sign in to write one
Any signed-in account that does not publish Proton Mail can review it, once — before installing it or after. A review from an account that has installed it is marked as one.
No reviews yet
Nobody has reviewed this listing. The rating on the card is the mean of the reviews written here and nothing else, so there is no rating until somebody writes the first — which takes a signed-in account that does not publish it, and nothing else.