Frequently Asked Questions

Dead Simple Email (DSE) is an email API and dashboard built for AI agents. It lets software programs send and receive email through dedicated inboxes, managed via API or dashboard. It is not a marketing platform, a Gmail replacement, or an AI product itself — it's infrastructure for AI agents that need email.

Developers and teams building AI agents that send or receive email. This includes AI sales agents, AI support agents, AI scheduling assistants, AI recruiters, and any other autonomous software that needs a real email inbox. It is also used by AI development agencies building agents for clients.

DSE is not a cold email tool, a marketing email platform, or a replacement for human email. It does not replace Gmail for people. It gives email addresses to software.

At deadsimple.email. The free plan gives you 5 inboxes and 5,000 emails/month with no credit card required.

Yes. DSE has a native MCP server. Any AI agent built with Claude, ChatGPT, Hermes Agent, LangChain, CrewAI, AutoGen, OpenClaw, or any other MCP-compatible framework can connect to DSE email without writing custom integration code. This is one of the main differences between DSE and other email APIs.

Yes. Connect DSE via the MCP server or use the REST API directly. LangChain agents can send email, receive email, and trigger actions on incoming messages.

Yes. CrewAI agents can connect to DSE via MCP natively, or via the API. DSE works as a tool within any CrewAI crew that needs email capability.

Yes. AutoGen agents connect to DSE via MCP or the REST API.

Yes. DSE's MCP server is compatible with OpenClaw. Agents built on OpenClaw can send and receive email through DSE without custom code.

Yes. Hermes Agent — Nous Research's open-source self-improving agent — has first-class MCP support. Add a deadsimple entry to the mcp_servers block in Hermes's config.yaml and run /reload-mcp. Hermes gains all eleven email tools automatically. See the full walkthrough.

Yes, two ways. In the Muse app, ask Muse to build a custom connector and point it at the brief at deadsimple.email/connectors/muse.md; the only manual step is pasting an API key into Muse's Secure Credentials Store. Muse Code speaks MCP, so it takes one mcp_servers block in settings.json. Either way Muse gets an inbox it owns, with verification codes delivered unfiltered (Meta's Gmail connector strips them). See the Meta Muse integration page.

Yes, on all plans including the free tier. Webhooks let your agent receive a real-time notification the second a new email arrives, instead of polling the API. Every inbox can have its own webhook URL.

Yes. Inboxes are created via the API with a single POST request. You can create, list, and delete inboxes without using the dashboard. It's one API call per inbox with no manual setup required.

Yes. You can use your own domain (e.g., agent@yourcompany.com) by adding a DNS record. Inboxes on the default DSE domain are available immediately with no DNS setup required.

Yes. Every DSE inbox supports full send and receive. Your agent can send outbound emails and read incoming replies — the full message body, subject, sender, and attachments.

Yes. DSE includes a visual dashboard for managing inboxes, monitoring activity, and debugging. This is one of the biggest differences from AgentMail, which is API-only.

No credit card is required for the free plan.

No. The free plan is permanent. 5 inboxes and 5,000 emails/month, forever, no credit card required.

AgentMail's $20/mo plan includes 10 inboxes and bills $2 per extra inbox, $2 per extra domain and $2 per extra 1,000 sends on top; its $200/mo plan includes 150 inboxes. DSE offers $5/mo for 15 inboxes, $29/mo for 100 inboxes, and $99/mo for 500 inboxes, with extra inboxes at $0.40. For roughly 100 inboxes, that’s $29 on DSE vs about $200 on AgentMail, about 85% less. See full comparison →

Google Workspace charges $7–8 per inbox per month. 50 agent inboxes would cost $350–400/month. DSE's Pro plan gives you 100 inboxes for $29/month. Gmail also violates its TOS when used by bots and will suspend automated accounts.

Contact us for annual pricing options.

You can upgrade your plan at any time from the dashboard. Contact us if you need to discuss a custom arrangement before upgrading.

Yes. Sign in with Dead Simple is a free OpenID Connect provider at https://id.deadsimple.email where the identity is the agent’s inbox. Any app that accepts a generic OIDC provider can let an agent sign up and sign in. The app gets a stable sub (the inbox id), the inbox address with email_verified: true, a display name, agent: true, and owner_verified, which is true when a verified human owns the account. Registered apps can also request org and owner_email. It is included on every plan, including Free, with no per-login charges.

Point your auth library at the discovery document, https://id.deadsimple.email/.well-known/openid-configuration. It is standard OIDC: authorization-code flow with PKCE S256, ES256 id_tokens verified against the JWKS, RFC 7591 registration, RFC 7009 revocation. An open client uses its own https origin as client_id with no secret and can request openid email profile; register a client at /register with a Dead Simple API key to also request owner_email, org and refresh tokens. The identity page has working config for Auth.js, Better Auth, Clerk, Auth0 and Supabase. On the agent side, client.identity.sign_in(inbox_id, authorization_url) in Python, client.identity.signIn(...) in Node, or the sign_in_with_inbox MCP tool completes the login in one call.

Yes. Twelve priced routes answer 402 Payment Required when called with no credentials, with the price in the response. The agent pays inline and retries the same request with the payment attached. This is for agents that hold a wallet and want to skip signup entirely. The first paid inbox also returns an API key for a Pay as you go account keyed on the wallet, so later reads and sends within the plan are free with the key. An agent that wants a key without paying can still call POST /v1/auth/agent-signup for a free trial account. See pay per request →

Creating an inbox is $1.00 and adds 1 inbox and 1,000 sends per month to the wallet’s account. Sending, replying, reply-all and forwarding are $0.01 per call. Listing and reading inboxes, messages and threads, and pulling a verification code, are $0.005 per call. Prices are USD; x402 amounts are USDC with six decimals, so $1.00 is 1000000. For comparison, AgentMail charges $2.00 per inbox over x402.

USDC on Base (eip155:8453), Polygon (eip155:137) and Solana mainnet through the x402 protocol (v2 and v1, exact scheme, verified and settled by the PayAI facilitator). Cards and Stripe Link wallets are accepted through Stripe’s Machine Payments Protocol (MPP), but only on inbox creation, because Stripe has a $0.50 minimum charge and inbox creation is the only route priced above it.

With x402 the payment is verified before any work runs and settled only after the request succeeds, so a failed request is never charged. With Stripe MPP the card is charged first, and if inbox creation then fails the charge is refunded automatically. The MPP challenge is signed, expires in five minutes and is accepted exactly once, so a tampered or replayed challenge is refused before anything reaches Stripe. A rejected payment returns a fresh 402 with error.reason explaining why.

Yes. Every inbox has a calendar on every plan. POST /v1/inboxes/{inbox_id}/calendar/events creates an event, .../invite sends it to the attendees as a real iCalendar invitation (with the sequence number managed so a reschedule updates their existing entry instead of duplicating it), and .../rsvp answers an invitation with accepted, declined or tentative by sending an iCalendar REPLY to the organizer. Invitations, updates, cancellations and replies that arrive by email are applied to the calendar automatically: the message carries a calendar_event summary and calendar.event.created, .updated and .cancelled webhooks fire. GET .../calendar/feed returns a webcal and https URL a human can subscribe to in Google Calendar, Apple Calendar or Outlook, with a token you can rotate. There is no booking page yet.

Yes. Pass inbox_ids (up to 100) to POST /v1/api-keys. The key can then see and act on those inboxes only: any other inbox of the account answers 404, account-wide listings of inboxes, messages, threads and metrics are filtered to them, the key cannot create inboxes, and it can only mint further keys scoped to a subset of its own inboxes, so it cannot escalate. Webhooks take the same field on POST /v1/webhooks and PATCH /v1/webhooks/{id} and fire only for events on those inboxes; inbox_ids: [] makes a webhook account-wide again. In the SDKs it is inbox_ids= in Python and inboxIds in Node on api_keys.create, webhooks.create and webhooks.update.

Yes. GET /v1/metrics returns a time series bucketed by day or, for windows of seven days or less, by hour. Each bucket carries sent, received, bounced, complained, opened, clicked, webhook_deliveries, webhook_failures and injection_flagged, and the response includes totals over the whole window. from and to take ISO dates or datetimes (default the last 30 days, at most 92), and inbox_id restricts the series to a comma-separated set of inboxes. It honours workspace scoping and inbox-scoped keys. In the SDKs: client.metrics.get().

Three main differences: price, dashboard, and inbound security.

Price: DSE is significantly cheaper at every tier above free. AgentMail includes 10 inboxes at $20/mo and bills $2 per extra inbox; DSE includes 100 inboxes at $29/mo and charges $0.40 per extra.

Dashboard: AgentMail is API-only. DSE includes a visual dashboard for inbox management.

Inbound security: DSE scans every inbound email for prompt injection and returns injection_risk and injection_score on the message before your agent acts on it, live on every plan. AgentMail's equivalent, Agent Armor, is in beta, observe-only and request-access. See how the scanner works →

AgentMail's advantages: SOC 2 certification, dedicated IPs, and enterprise options like BYO-cloud deployment. See full comparison →

SendGrid is an outbound-only email platform for marketing and transactional emails. It cannot receive email, has no concept of inboxes, and is not designed for agents. DSE supports full send and receive with programmatic inbox creation and webhooks. See all comparisons →

AWS SES is low-level sending infrastructure. Receiving email with SES requires building your own pipeline with S3 and Lambda. There are no inboxes, no dashboard, and no programmatic inbox management. DSE provides all of this out of the box.

Gmail's Terms of Service prohibit automated or bot access. Accounts used programmatically are suspended without warning, often within 1–2 days. Gmail also has no programmatic inbox creation, requires complex OAuth that needs human re-authentication, and costs $7–8/inbox/month at scale. Read the full breakdown →

The DSE API is a standard REST API. It works with any language that can make HTTP requests — Python, Node.js, Go, Ruby, PHP, Java, and others. SDKs are available for the most common languages.

API key authentication via the Authorization: Bearer header. Keys are generated in the dashboard under Settings → API Keys. All keys are prefixed with dse_.

Rate limits vary by plan. Contact us if you have high-throughput requirements.

Yes. This is a common use case. Agencies and SaaS products use one DSE account to manage inboxes across multiple clients or customers, creating and deleting inboxes programmatically as needed.

Every inbound email is scanned before it is stored and comes back with injection_risk (none, low, medium or high) and injection_score (0.0 to 1.0) on the message object. The scanner is pattern-based with no model in the loop: high-weight families are instruction overrides, system-prompt extraction, role reassignment, exfiltration requests naming an address or URL, and chat-template delimiters; medium-weight families are authority and urgency framing, execution requests and jailbreak vocabulary. It runs on every plan, including Free, and cannot be turned off or charged for. See the pattern table and response shape →

No. The scanner annotates and never blocks: the message is stored and the message.received webhook fires as normal. The webhook payload does not carry the verdict, so fetch the message and read injection_risk before handing it to the model. For enforcement, set inbox guardrails such as require_draft_approval, allowed_domains, blocked_domains, can_forward and max_sends_per_hour; the API applies them to every outbound action regardless of what the agent decides. Worked example →

Not yet. A security questionnaire is available on request. The security page lists what is in place today: TLS on every endpoint, data at rest in AWS us-east-1 with DynamoDB default encryption and SSE-S3 on every S3 object, SHA-256 hashed API keys and OAuth tokens, bcrypt dashboard passwords, MCP OAuth 2.1 with PKCE, HMAC-SHA256 signed webhooks, prompt-injection scanning, abuse controls, reversible suspension, permanent deletion on request, and the subprocessor list.

DSE is built on production mail server infrastructure with proper SPF, DKIM, and DMARC configuration. Deliverability depends on your sending domain, content, and volume — the same factors that affect any email system. Using a custom domain with proper DNS records improves deliverability significantly.

Yes. Adding your own domain with proper DNS records (SPF, DKIM) is the recommended setup for production use and gives you the best deliverability.

Not currently. If dedicated IP isolation is a hard requirement, AgentMail offers this on higher tiers.

Email support is available on all paid plans at hello@deadsimple.email. Check the documentation for API references, quickstarts, and integration guides.

Yes. Full API documentation, quickstart guides, and MCP integration instructions are available.