# MultiMail

> Email infrastructure for AI agents with graduated oversight, verifiable identity, signed disclosure headers, pre-send domain intelligence, and MCP tooling.

Version: `0.9.0`
Date: `May 2026`

## Agent Quickstart

### Get email tools now (one line)

Connect to the hosted MCP server — no install, no API key needed to start:

```json
{ "mcpServers": { "multimail": { "type": "url", "url": "https://mcp.multimail.dev/mcp" } } }
```

Then create your free account using the tools the server exposes:

```
request_challenge → create_account → activate_account → configure_mailbox → send_email
```

That's it. Five tool calls from zero to sending email. No npm, no pip, no API key up front — `request_challenge` and `create_account` work without authentication. Your API key (`mm_live_...`) comes back from `create_account`. Your mailbox address is `agent@{slug}.multimail.dev`.

### What each tool does in the signup flow

| Step | Tool | Auth required | What it does |
|------|------|---------------|-------------|
| 1 | `request_challenge` | No | Returns a challenge to solve |
| 2 | `create_account` | No | Submit solved challenge + operator name, email, slug. Returns API key + mailbox ID |
| 3 | `activate_account` | Yes | Redeem the activation code sent to operator email |
| 4 | `configure_mailbox` | Yes | Set oversight mode, display name, signature (optional — defaults work) |
| 5 | `send_email` | Yes | Send markdown email. Converts to HTML automatically |

### No MCP? Use the REST API directly

The hosted MCP server is OAuth-gated (it expects a browser login), so a headless agent with only HTTP (curl/fetch) can't connect it. Do the same signup over plain HTTPS instead — `request_challenge` → `create_account` → `activate_account` map 1:1 to these REST calls. No API key is needed to start. (Call the API with curl or a normal User-Agent — Cloudflare returns 403 to a bare `urllib`/default-library UA.)

1. **Request a challenge:** `POST https://api.multimail.dev/v1/account/challenge` → `{"algorithm":"SHA-256","challenge":"<hex>","salt":"<salt>","maxnumber":<N>,"signature":"<sig>"}`

2. **Solve the proof-of-work:** find the integer `number` in `[0, maxnumber]` where the hex `SHA-256(salt + number)` equals `challenge`. A brute-force loop finds it in well under a second.

3. **Create the account** — pass the solved challenge as a `pow_solution` **object** (the five fields below, with your `number`), *not* a base64 string:

```
curl -sX POST https://api.multimail.dev/v1/account \
  -H "Content-Type: application/json" -d '{
    "operator_name": "Greenline Studios",
    "oversight_email": "ops@greenline.studio",
    "slug": "greenline",
    "accepted_tos": true,
    "accepted_operator_agreement": true,
    "accepted_anti_spam_policy": true,
    "email_use_type": "transactional",
    "pow_solution": {"algorithm":"SHA-256","challenge":"<hex>","number":<N>,"salt":"<salt>","signature":"<sig>"}
  }'
```

Response: `200 {"status":"confirmation_sent"}`. This emails a one-time activation code to your operator (`oversight_email`). **No API key is returned at this step.**

4. **Ask your operator for the activation code.** MultiMail sends it to the human's inbox — you can't read it yourself. Ask them to paste it back to you. (By design: a human approves every new agent account.)

5. **Confirm to receive your API key:** `POST https://api.multimail.dev/v1/confirm` with `Content-Type: application/json` and `{"code":"ABC-XK7-9M2"}` → `200 {"api_key":"mm_live_...","mailbox_id":"...","mailbox_address":"agent@greenline.multimail.dev"}`. Save the `api_key` — it's shown once. Use it as `Authorization: Bearer mm_live_...` from here on.

6. **Send** (default `gated_send`, so your operator approves each outbound until you've earned more autonomy): `POST https://api.multimail.dev/v1/mailboxes/<mailbox_id>/send` with the Bearer header and your email body. See the OpenAPI spec for the full schema.

### Local install (alternative)

```json
{
  "mcpServers": {
    "multimail": {
      "command": "npx",
      "args": ["-y", "@multimail/mcp-server"],
      "env": { "MULTIMAIL_API_KEY": "mm_live_..." }
    }
  }
}
```

### What happens when you send

Default oversight mode is `gated_send`: your `send_email` call returns `status: "pending"` and the operator gets an approval email. Once approved, the email delivers.

To send without approval, set the mailbox to `monitored` or `autonomous` via `configure_mailbox`. Five modes:

- `read_only` — agent reads only, sends blocked
- `gated_all` — inbound and outbound require approval
- `gated_send` — inbound auto, outbound requires approval **(default)**
- `monitored` — sends automatically, operator gets a copy
- `autonomous` — full send/receive without approval

### Free tier

2 mailboxes · 200 emails/month · 100 MB storage

### Key resources

- OpenAPI spec: `https://api.multimail.dev/v1/openapi.json`
- Full LLM docs: `https://multimail.dev/llms-full.txt`
- API quickstart (curl): `https://multimail.dev/quickstart-api`
- MCP quickstart: `https://multimail.dev/quickstart-mcp`

---

## Overview

MultiMail is an email API and MCP server for AI agents.

Core runtime:
- Serverless edge runtime
- A metadata store
- Object storage for email/attachment blobs
- A key-value cache for caching, rate limits, approvals, and reputation data
- Async job queues for threat scanning, scheduling, provisioning, and webhook delivery
- A stateful scheduler for delayed sends

What MultiMail provides:
- Agent mailboxes
- Inbound email ingestion to markdown
- Outbound markdown-to-HTML delivery
- Reply and thread tracking
- Human oversight modes
- Signed identity headers
- Domain intelligence before send
- Content filtering and attachment blocking
- Rate limiting and mailbox warmup
- Reputation and enforcement
- Contacts, tags, webhooks, suppressions, audit log, usage, export
- 51 MCP tools

## Key Concepts

### Oversight Modes

- `read_only`: inbound read access only; outbound blocked
- `gated_all`: inbound and outbound require approval
- `gated_send`: inbound automatic; outbound requires approval
- `monitored`: sends automatically; operator receives monitoring copy
- `autonomous`: sends automatically without approval

Default mailbox mode:
- `gated_send`

Upgrade rule:
- Normal mailbox updates can only downgrade oversight
- Higher-autonomy upgrades require `request-upgrade` then `upgrade` with operator code

Enforcement override:
- `restricted` tenant enforcement forces all outbound through approval and lowers send rate limits
- `suspended` and `terminated` disable sending regardless of mailbox mode

### Agent Identity

Outbound mail can include:
- `X-MultiMail-Identity`: signed ECDSA P-256 identity claim
- `X-AI-Generated`: convenience disclosure header when AI disclosure is enabled
- `X-MultiMail-Reputation`: mailbox reputation hash

Signed `IdentityClaim` fields:
- `ai_generated`
- `operator`
- `oversight`
- `capabilities`
- `verified_operator`
- `created`
- `service`
- `iat`
- `reputation_hash`

Signing details:
- ECDSA P-256
- SHA-256
- Canonical JSON with sorted keys
- Public key published at `/.well-known/multimail-signing-key`

Mailbox setting:
- AI disclosure is always on — enforced at the send gateway (EU AI Act Article 50); `ai_disclosure: false` is rejected with 400

Oversight disclosure text used in signatures:
- `read_only`: This AI agent can only read emails
- `gated_all`: All actions by this AI agent require human approval
- `gated_send`: Emails from this AI agent are approved by a human before sending
- `monitored`: This AI agent's emails are monitored by a human
- `autonomous`: This AI agent sends emails independently

### Domain Intelligence

MultiMail checks recipient domains before send.

Signals used:
- MX existence
- MX provider fingerprinting
- SPF presence
- DMARC presence and policy
- SOA serial freshness
- Spamhaus DBL listing
- Disposable domain list
- Platform bounce history for that domain

Classification:
- `green`: all checks passed
- `yellow`: degraded but not blocked
- `red`: blocked before send

Red examples:
- disposable domain
- no MX
- Spamhaus listed
- parked-domain MX provider
- hard bounce rate over 30% with at least 10 sends

Yellow examples:
- DMARC `p=none`
- bounce rate over 10% with at least 10 sends
- stale SOA serial
- self-hosted MX without SPF

Caching:
- green 24h
- yellow 6h
- red 6h to 7d depending on reason
- fail-open if DNS is unreachable

### Content Filtering

Outbound content filtering runs before send.

Blocks include:
- empty body
- too many recipients
- body larger than 256 KB
- subject longer than 998 chars
- spam patterns
- header injection
- invalid recipient addresses
- malware patterns
- brand impersonation
- promotional/marketing content
- URL shorteners

Attachment restrictions:
- executable/script extensions like `exe`, `bat`, `cmd`, `scr`, `ps1`, `js`, `vbs`, `msi`, `dll`, `sys`
- dangerous Windows executable MIME types
- double-extension decoys like `invoice.pdf.exe`

Platform policy:
- transactional email only
- not for bulk or promotional marketing

### Reputation

Public mailbox reputation is exposed through a hashed address identity.

Public endpoint:
- `GET /.well-known/reputation/{hash}`

Reputation-related behaviors in code:
- hard bounces add permanent suppression entries
- spam complaints add suppression entries
- bounce and complaint rates feed enforcement

### Scheduled Send

Supported request fields:
- `send_at`
- `gate_timing`

Gate timing modes:
- `gate_first`: approve before scheduling
- `schedule_first`: schedule first, approve at delivery time

Plan-based schedule windows:
- Starter: up to 1 day ahead
- Builder: up to 7 days ahead
- Pro: up to 30 days ahead
- Scale: up to 30 days ahead

### Presigned Attachments

Attachment access:
- inline download for small files
- signed temporary URL for larger files

Public attachment token endpoint:
- `GET /v1/attachments/{token}`

Signed URL lifetime:
- 1 hour

## Compliance

### AI Disclosure and Provenance

- EU AI Act Article 50 support via signed `ai_generated` claim and optional `X-AI-Generated` header
- Machine-readable provenance via `X-MultiMail-Identity`
- Lean 4 proofs cover tamper evidence for signed identity headers
- Proof status endpoint: `GET /v1/proof-status`

### US State AI Disclosure Positioning

Current product materials reference built-in support for:
- Maine `LD 1727` (September 2025)
- New York `S-3008C` (November 2025)
- California `SB 243` (January 2026)
- Illinois state AI disclosure requirements (January 2026)

### Regulatory Readiness (EU Agent Identity)

Full mapping of MultiMail's trust chain to eIDAS 2.0, EU AI Act, Apply AI Strategy, and Product Liability Directive requirements: [multimail.dev/regulatory](https://multimail.dev/regulatory)

The four-layer trust chain (DKIM → X-MultiMail-Identity → ECDSA P-256 → .well-known public key) is structurally a did:web agent identity binding — the compliance primitive CEPS named as missing from European regulation (May 2026).

### Agent Registration (auth.md)

MultiMail supports the [auth.md](https://github.com/workos/auth.md) open protocol for agent registration. Auth.md is a Markdown file hosted at a domain (e.g., `https://service.com/auth.md`) that describes how agents register, which OAuth flows are supported, and available scopes. It composes RFC 9728 (Protected Resource Metadata), ID-JAG identity assertions, and OIDC backchannel logout.

MultiMail's existing agent signup flow (`request_challenge` → `create_account` → `activate_account`) maps to auth.md's "user claimed" flow (OTP-based registration). The trust ladder (read_only → gated_send → monitored → autonomous) provides the graduated authorization that auth.md's scope model expects.

Spec: [workos.com/auth-md/docs](https://workos.com/auth-md/docs)

### Other Compliance Positioning

- CAN-SPAM support through unsubscribe tokens and physical-address support
- NIST agent identity positioning through interoperable, cryptographic, domain-anchored email identity
- EU Product Liability Directive positioning for auditable agent actions
- Insurability positioning through oversight, signed provenance, and audit trails

## Authentication

Actual auth format in code:

```text
Authorization: Bearer $MULTIMAIL_API_KEY
```

API key format:
- prefix: `mm_live_`
- 32 random bytes, base64url encoded

Scopes:
- `read`
- `send`
- `admin`
- `oversight`
- `oversight_read`

Scope rule:
- `admin` implies all scoped checks in the worker

Account activation:
- new accounts start inactive
- operator receives activation code
- operator redeems code at `/v1/confirm`
- starter signup returns API key only after activation

Admin action approval flow:
- mailbox creation
- API key creation
- webhook creation

These require:
1. initial request without `approval_code`
2. code emailed to operator
3. repeat request with `approval_code`

## API

Base URL:
- `https://api.multimail.dev`

OpenAPI:
- `https://api.multimail.dev/v1/openapi.json`

Public key:
- `https://api.multimail.dev/.well-known/multimail-signing-key`

Proof status:
- `https://api.multimail.dev/v1/proof-status`

### Public and Signup Endpoints

- `GET /v1/slug-check/:slug` - Check tenant slug availability and suggestions
- `POST /v1/account` - Create starter account and default mailbox
- `POST /v1/account/resend-confirmation` - Resend operator activation code
- `GET /v1/setup/verify-token` - Validate onboarding setup token
- `GET /v1/confirm` - Render activation code form
- `GET /v1/confirm/:code` - Redirect with prefilled activation code
- `POST /v1/confirm` - Activate account with code
- `GET /v1/unsubscribe/:token` - Render unsubscribe page
- `POST /v1/unsubscribe/:token` - One-click/browser unsubscribe
- `GET /v1/approve/:token` - Hosted approval page or one-click approve/reject
- `POST /v1/approve/:token` - Submit hosted approval decision
- `GET /v1/openapi.json` - OpenAPI spec
- `POST /v1/account` - Create account
- `GET /v1/proof-status` - Lean proof verification status
- `GET /health` - Service health

### Authenticated Account and Mailbox Endpoints

- `GET /v1/account` - Get account info, plan, quota, storage, sending status
- `PATCH /v1/account` - Update account name, oversight email, physical address
- `DELETE /v1/account` - Permanently delete account with confirmation body
- `POST /v1/mailboxes` - Create mailbox after operator approval
- `GET /v1/mailboxes` - List mailboxes
- `PATCH /v1/mailboxes/:mailboxId` - Update mailbox settings and downgrade oversight
- `PATCH /v1/mailboxes/:mailboxId/configure` - MCP onboarding mailbox configuration
- `DELETE /v1/mailboxes/:mailboxId` - Soft-delete mailbox
- `POST /v1/mailboxes/:mailboxId/request-upgrade` - Request higher oversight mode
- `POST /v1/mailboxes/:mailboxId/upgrade` - Redeem oversight upgrade code

### Email and Thread Endpoints

- `GET /v1/mailboxes/:mailboxId/emails` - List email summaries with filters
- `GET /v1/mailboxes/:mailboxId/emails/:emailId` - Get full email and markdown body
- `GET /v1/mailboxes/:mailboxId/emails/:emailId/attachments/:filename` - Download attachment
- `GET /v1/mailboxes/:mailboxId/emails/:emailId/attachments/:filename/url` - Get signed attachment URL
- `GET /v1/attachments/:token` - Public download via signed token
- `POST /v1/mailboxes/:mailboxId/send` - Send or schedule outbound email
- `POST /v1/mailboxes/:mailboxId/reply/:emailId` - Reply in existing thread
- `GET /v1/mailboxes/:mailboxId/threads/:threadId` - Get thread summary and messages
- `POST /v1/mailboxes/:mailboxId/emails/:emailId/cancel` - Cancel pending or scheduled email
- `PATCH /v1/mailboxes/:mailboxId/emails/:emailId/schedule` - Edit scheduled email
- `PUT /v1/mailboxes/:mailboxId/emails/:emailId/tags` - Set email tags
- `GET /v1/mailboxes/:mailboxId/emails/:emailId/tags` - Get email tags
- `DELETE /v1/mailboxes/:mailboxId/emails/:emailId/tags/:key` - Delete one tag

### Contacts, Oversight, Suppression, Audit, Usage, Export

- `POST /v1/contacts` - Create contact
- `GET /v1/contacts` - Search or list contacts
- `DELETE /v1/contacts/:contactId` - Delete contact
- `GET /v1/oversight/pending` - List pending oversight emails
- `POST /v1/oversight/decide` - Approve or reject pending email
- `GET /v1/suppression` - List suppressed addresses
- `DELETE /v1/suppression/:address` - Remove non-permanent suppression
- `GET /v1/audit-log` - List tenant audit log
- `GET /v1/usage` - Usage summary or daily breakdown
- `GET /v1/export` - Export tenant data bundle

### API Keys

- `GET /v1/api-keys` - List API keys
- `POST /v1/api-keys` - Create API key after operator approval
- `PATCH /v1/api-keys/:keyId` - Update API key metadata/scopes
- `DELETE /v1/api-keys/:keyId` - Revoke API key

### Custom Domains

- `POST /v1/domains` - Register custom sending domain
- `GET /v1/domains` - List custom domains
- `GET /v1/domains/:id` - Get custom domain detail
- `POST /v1/domains/:id/verify` - Verify DNS records
- `DELETE /v1/domains/:id` - Delete custom domain

### Webhooks and Delivery History

- `POST /v1/webhooks` - Create webhook after operator approval
- `GET /v1/webhooks` - List webhook subscriptions
- `GET /v1/webhooks/:id` - Get webhook detail including signing secret
- `DELETE /v1/webhooks/:id` - Delete webhook subscription
- `GET /v1/webhook-deliveries` - List webhook delivery attempts

### Billing

- `POST /v1/billing/checkout` - Authenticated Stripe checkout
- `POST /v1/billing/pricing-checkout` - Public pricing-page Stripe checkout
- `GET /v1/billing/session-key` - One-time API key retrieval after pricing checkout
- `POST /v1/billing/stripe-webhook` - Stripe webhook
- `POST /v1/billing/crypto-checkout` - Coinbase Commerce checkout
- `POST /v1/billing/coinbase-webhook` - Coinbase webhook

### Well-Known Reputation and Key Discovery

- `GET /.well-known/multimail-signing-key` - Public verification key
- `GET /.well-known/reputation/:hash` - Public mailbox reputation by hash

### Admin Endpoints

- `GET /v1/admin/tenants/:id/enforcement` - Get tenant enforcement state
- `POST /v1/admin/tenants/:id/reactivate` - Clear enforcement
- `POST /v1/admin/tenants/:id/suspend` - Set `restricted` or `suspended`
- `POST /v1/admin/tenants/:id/terminate` - Terminate tenant
- `GET /v1/admin/enforcement/pending` - List enforced tenants
- `GET /v1/admin/verify` - Verify admin key
- `GET /v1/admin/tenants` - List tenants
- `GET /v1/admin/audit-log` - Admin audit log

## Email Statuses

- `unread`
- `read`
- `archived`
- `deleted`
- `pending_send_approval`
- `pending_inbound_approval`
- `pending_scan`
- `scan_blocked`
- `rejected`
- `cancelled`
- `send_failed`
- `scheduled`
- `blocked_domain`

## Rate Limits

Per API key:
- read: 60 requests/minute
- send: 10 requests/minute
- restricted send: 2 requests/minute
- upgrade request: 3/hour per mailbox key bucket
- export: 1/hour per tenant

Mailbox warmup daily caps:
- day 1-3: starter 10/day, paid 20/day
- day 4-7: starter 20/day, paid 40/day
- day 8-14: starter 40/day, paid 80/day
- day 15-30: starter 80/day, paid 160/day
- after day 30: warmed up

## Enforcement

Enforcement tiers:
- `none`
- `warning`
- `restricted`
- `suspended`
- `terminated`

Triggers in code:
- bounce warning above 2% with enough volume
- restricted for bounce rate above 5%
- restricted for spam complaint rate above 0.3% with enough delivered volume
- suspended for threat detection
- suspended for unusual bulk sending
- terminated by admin action

Effects:
- `warning`: notification and audit trail
- `restricted`: mandatory approval and reduced send rate
- `suspended`: sending disabled
- `terminated`: sending disabled and tenant deactivated

## MCP Server

Install:

```json
{
  "mcpServers": {
    "multimail": {
      "command": "npx",
      "args": ["-y", "@multimail/mcp-server"],
      "env": {
        "MULTIMAIL_API_KEY": "$MULTIMAIL_API_KEY"
      }
    }
  }
}
```

Environment variables:
- `MULTIMAIL_API_KEY`
- optional `MULTIMAIL_MAILBOX_ID`
- optional `MULTIMAIL_API_URL`

51 tools:
- `request_challenge` - Get a signup challenge (no API key required)
- `request_did_challenge` - Get a DID identity-binding challenge for optional verifiable-identity signup (no API key required)
- `create_account` - Create account with solved challenge, optionally binding an Ed25519 agent DID (no API key required)
- `activate_account` - Activate account with confirmation code
- `list_mailboxes` - List available mailboxes
- `create_mailbox` - Create mailbox with approval code flow
- `update_mailbox` - Update mailbox settings
- `configure_mailbox` - Configure oversight mode, auto-BCC, and mailbox behavior
- `delete_mailbox` - Delete mailbox
- `send_email` - Send markdown email with scan/approval awareness
- `reply_email` - Reply in-thread
- `check_inbox` - List inbox email summaries with filters
- `read_email` - Read full email and markdown body
- `download_attachment` - Download attachment inline or by signed URL
- `get_thread` - Get thread participants and message list
- `edit_scheduled_email` - Edit scheduled email before send
- `cancel_message` - Cancel pending or scheduled message
- `get_tags` - Get all tags on an email
- `set_tags` - Set key-value tags on an email (merge with existing)
- `delete_tag` - Delete a specific tag key from an email
- `search_contacts` - Search contacts by name or email
- `add_contact` - Add a contact
- `delete_contact` - Delete a contact
- `list_pending` - List pending oversight items
- `decide_email` - Approve or reject pending email
- `list_spam` - List spam messages
- `manage_spam_status` - Report spam or clear spam status (action: report|clear)
- `list_suppression` - List suppressed recipients with pagination
- `remove_suppression` - Remove a recipient from the suppression list
- `get_account` - Inspect plan, quota, and enforcement state
- `update_account` - Update account settings
- `delete_account` - Delete account
- `resend_confirmation` - Resend activation code
- `get_usage` - Get usage and plan-limit data
- `list_api_keys` - List API keys
- `create_api_key` - Create API key with approval flow
- `revoke_api_key` - Revoke API key
- `get_audit_log` - Fetch audit history
- `wait_for_email` - Poll efficiently until matching mail arrives
- `create_webhook` - Create webhook with approval flow
- `list_webhooks` - List all webhooks
- `delete_webhook` - Delete a webhook
- `list_allowlist` - List sending allowlist entries
- `add_allowlist_entry` - Add a recipient to the sending allowlist
- `remove_allowlist_entry` - Remove an allowlist entry
- `manage_upgrade` - Request or apply oversight upgrade (action: request|apply)
- `upgrade_plan` - Upgrade to a paid Stripe plan (Builder, Pro, Scale)
- `cancel_subscription` - Cancel paid Stripe subscription, revert to starter at period end
- `get_billing_portal` - Generate a Stripe customer-portal URL for self-service billing
- `report_issue` - Report a tool bug, site problem, or feature request
- `setup_multimail` - Onboarding tool on the public /onboard endpoint; no API key required

## Pricing

### Starter
- price: free
- mailboxes: 2
- monthly quota: 200 emails
- storage: 100 MB
- schedule window: 1 day ahead
- scheduled emails per mailbox: 10

### Builder
- price: $9/month
- mailboxes: 5
- monthly quota: 5,000 emails
- storage: 10 GB
- schedule window: 7 days ahead
- scheduled emails per mailbox: 50

### Pro
- price: $29/month
- mailboxes: 25
- monthly quota: 30,000 emails
- storage: 50 GB
- schedule window: 30 days ahead
- scheduled emails per mailbox: 200
- custom domains: yes

### Scale
- price: $99/month
- mailboxes: 100
- monthly quota: 150,000 emails
- storage: 150 GB
- schedule window: 30 days ahead
- scheduled emails per mailbox: 1,000
- custom domains: yes

## Use Cases (144)

### Developer (32)
- `ci-cd-notifications`, `deployment-alerts`, `deployment-notifications`, `pr-review-requests`, `incident-postmortems`, `api-key-rotation`, `audit-log-reports`, `knowledge-base-updates`, `api-usage-alerts`, `email-api-for-ai-agents`, `inbound-email-parsing`, `agent-to-agent-email`, `email-inbox-api`, `resend-alternative`, `postmark-alternative`, `sendgrid-alternative`, `mailgun-alternative`, `agentmail-alternative`, `how-to-send-email-with-ai-agent`, `how-to-build-ai-email-agent`, `how-to-verify-ai-email-sender`, `how-to-set-up-email-oversight`, `how-to-handle-inbound-email-with-ai`, `email-webhooks-for-ai-agents`, `email-attachments-api`, `email-threading-api`, `email-contacts-api`, `email-templates-for-ai-agents`, `ai-email-for-devops`, `dkim-for-ai-agents`, `email-authentication-for-ai`, `mcp-email-server`

### Finance (8)
- `invoice-follow-up`, `invoice-processing`, `payment-reminders`, `expense-approvals`, `receipt-processing`, `collection-notices`, `investor-updates`, `donor-acknowledgments`

### HR (8)
- `interview-scheduling`, `offer-letter-delivery`, `onboarding-checklists`, `benefits-enrollment`, `exit-surveys`, `employee-recognition`, `candidate-rejection`, `compliance-training-reminders`

### Legal (17)
- `contract-reminders`, `compliance-notifications`, `nda-follow-ups`, `regulatory-updates`, `case-updates`, `data-breach-notification`, `eu-ai-act-email-compliance`, `ai-disclosure-headers`, `state-ai-disclosure-laws`, `agent-insurability`, `agent-identity-standards`, `eu-product-liability-ai-agents`, `gdpr-ai-email-compliance`, `ccpa-ai-email-disclosures`, `ai-email-for-cisos`, `ai-email-for-compliance-officers`, `verifiable-intent-agent-commerce`

### Marketing (11)
- `newsletter-curation`, `event-invitations`, `product-announcements`, `feedback-requests`, `referral-programs`, `survey-distribution`, `testimonial-requests`, `drip-campaigns`, `webinar-follow-up`, `seasonal-campaigns`, `fundraising-campaigns`

### Notifications (12)
- `system-alerts`, `usage-reports`, `billing-reminders`, `onboarding-emails`, `password-resets`, `account-verification`, `emergency-notifications`, `security-awareness`, `lab-result-notifications`, `prescription-refill-alerts`, `real-time-alerts`, `magic-link-emails`

### Operations (29)
- `order-confirmations`, `shipping-updates`, `inventory-alerts`, `vendor-communication`, `incident-reports`, `appointment-reminders`, `delivery-confirmations`, `return-status-updates`, `subscription-management`, `vendor-onboarding`, `customer-winback-notification`, `internal-announcements`, `customer-migration`, `sla-breach-alerts`, `content-moderation-notices`, `supplier-quality-reports`, `price-change-notifications`, `tenant-communications`, `weather-dependent-notifications`, `transactional-email-for-ai-agents`, `email-classification`, `email-routing`, `bounce-management`, `email-deliverability-optimization`, `email-sentiment-analysis`, `email-verification-for-ai-agents`, `email-oversight-modes`, `scheduled-email-for-ai-agents`, `ai-email-for-ctos`

### Personal (7)
- `email-summarization`, `meeting-prep`, `travel-booking`, `reminder-emails`, `birthday-anniversary-reminders`, `email-drafting`, `meeting-notes-distribution`

### Sales (10)
- `lead-qualification`, `sales-follow-up`, `meeting-scheduling`, `proposal-sending`, `win-back-campaigns`, `cold-outreach`, `competitive-intel`, `partner-communications`, `renewal-upsell`, `abandoned-cart-recovery`

### Support (10)
- `customer-support-auto-reply`, `ticket-triage`, `escalation-routing`, `faq-responses`, `satisfaction-surveys`, `customer-health-checks`, `warranty-claims`, `multi-language-support`, `customer-education`, `auto-responder-for-ai-agents`

## Industries (48)

- `healthcare`, `finance-banking`, `insurance`, `pharma`, `legal`, `real-estate`, `education-k12`, `higher-education`, `government`, `defense-military`, `retail-ecommerce`, `saas-technology`, `telecommunications`, `energy-utilities`, `manufacturing`, `construction`, `transportation-logistics`, `hospitality-travel`, `food-beverage`, `nonprofit`, `media-publishing`, `advertising-marketing`, `accounting-tax`, `hr-recruiting`, `cybersecurity`, `automotive`, `agriculture`, `aerospace`, `entertainment-gaming`, `professional-services`, `fintech`, `private-equity`, `venture-capital`, `recruiting-staffing`, `biotech`, `martech`, `logistics-supply-chain`, `edtech`, `crypto-web3`, `sports-entertainment`, `telehealth`, `cannabis`, `automotive-ev`, `space-aerospace`, `food-delivery`, `coworking-flex`, `pet-care`, `fitness-wellness`

## Framework Integrations (84)

- `langchain`, `langgraph`, `crewai`, `autogen`, `llamaindex`, `vercel-ai-sdk`, `openai-agents-sdk`, `semantic-kernel`, `haystack`, `dspy`, `smolagents`, `pydantic-ai`, `instructor`, `marvin`, `magentic`, `controlflow`, `langflow`, `flowise`, `n8n`, `rivet`, `babyagi`, `metagpt`, `chatdev`, `camel`, `letta`, `autogen-studio`, `taskweaver`, `composio`, `e2b`, `modal`, `dust`, `anthropic-api`, `openai-api`, `google-gemini-api`, `mistral-api`, `groq-api`, `together-ai-api`, `fireworks-api`, `aws-bedrock`, `azure-ai`, `vertex-ai`, `ollama`, `lm-studio`, `jan`, `superagent`, `agentgpt`, `phidata`, `ag2`, `google-adk`, `mastra`, `agno`, `livekit-agents`, `trigger-dev`, `inngest`, `salesforce`, `hubspot`, `zendesk`, `intercom`, `slack`, `browser-use`, `temporal-ai`, `prefect-ai`, `anthropic-sdk`, `openai-assistants-api`, `cohere`, `ai21`, `replicate`, `huggingface`, `make`, `zapier-ai`, `retool`, `superblocks`, `discord`, `notion`, `linear`, `github-actions`, `microsoft-agent-365`, `servicenow-ai-gateway`, `stripe-projects-dev`, `cloudflare-workers`, `fastapi`, `fixie`, `swarm`, `gmail`

## MCP Clients (23)

- `claude-desktop`, `claude-code`, `cursor`, `windsurf`, `vscode-copilot`, `vscode-continue`, `vscode-cline`, `zed`, `jetbrains`, `neovim`, `emacs`, `warp`, `raycast`, `obsidian`, `librechat`, `open-webui`, `jan-ai`, `msty`, `smithery`, `glama`, `chatgpt`, `gemini-cli`, `amazon-q-developer`

## Links

- Website: `https://multimail.dev`
- Docs: `https://multimail.dev/docs`
- Quickstart API: `https://multimail.dev/quickstart-api`
- Quickstart MCP: `https://multimail.dev/quickstart-mcp`
- Examples: `https://multimail.dev/examples`
- Error reference: `https://multimail.dev/errors`
- Patterns: `https://multimail.dev/patterns`
- Status: `https://multimail.dev/status`
- GitHub: `https://github.com/multimail-dev/mcp-server`
- Full LLM docs: `https://multimail.dev/llms-full.txt`