API and AI assistants

The same actions, from anywhere

Everything you do in the dashboard also works from code, through the REST API, and from AI assistants, through the MCP server: writing drafts, sending tests, scheduling, adding contacts, building segments, reading reports. Each action runs the same checks and limits as in the dashboard. The API and MCP docs have the setup and every endpoint.

For the unified audience, use Subscriber actions to read one person and their separate newsletter subscriptions. Contact IDs remain compatible as subscription IDs; old Recipient reads are for accounts awaiting migration. Group actions need a newsletter parent and never grant consent. The CSV preview is read-only, and the existing importer is separate from the later durable CSV and MailerLite Classic importers.

For purchase integrations, an authorized POST /api/v1/contacts call can pass subscribe: true with subscriber read/write and send permission (or a legacy full key). On a ready unified audience this records the owner's authorization to receive that newsletter without recipient confirmation. The owner chooses the audience; withdrawals, dormant status and delivery blocks remain in force. The response acknowledges subscriptionMode: owner-authorized; check status for any preserved restriction. Without this option the contact is organizational only. No welcome or confirmation email is sent.

AI assistants

ChatGPT, Claude and Grok Bot connect with your sign-in: add ApresMail in the assistant, then approve it on ApresMail's page. Claude Code, Codex and Cursor connect with an API key. Your assistant can use only the permissions you choose. Drafting does not permit tests, scheduled sends or subscriber reads. Sending to your list can't be undone, so it asks you first. In ChatGPT, OpenAI lets apps make changes only on its Business, Enterprise and Edu plans. See the setup for each assistant.

Apps connected with your sign-in are listed on the API keys page with their permissions and last use. Older connections may retain legacy full access. Revoke and reconnect them to choose a smaller grant.

The email on the approval page is the account the assistant will access. If you entered from the shared dashboard, connecting an app may ask you to sign in again. Use Switch to choose a different account for this connection; it does not sign you out of the shared dashboard.

API keys

Create keys on the API keys page. A key is shown once: copy it then, since ApresMail only keeps a fingerprint of it.

• Content: read content and aggregate results, or create and edit drafts.
• Sending: send tests, approve or schedule campaigns, publish or resume sequences, and enable signup or sending flows. Adding someone to a group can start a sequence and also needs this permission.
• Subscriber data: read people, exports and individual history separately from managing subscribers. Some writes return personal data and need both permissions.
• Settings and publication: configure newsletters, sending domains and forms. Live sending settings and form publication also need sending permission.

New keys expire after 30 days by default; choose 7, 30 or 90 days. Sending and subscriber access start unchecked. Scoped agents cannot create more keys or change their own permissions. Keep every key private. Older full and signups keys retain their existing behavior; the page labels them as legacy.

An account has up to 20 active keys. Revoking one stops it at once, for good. The authentication page lists the rate limits.