canonical: https://jentic.com/apis/agentboxd.com/agentboxd

# Agentboxd API

The Agentboxd API gives AI agents email inboxes, agent-to-agent messaging, and sign-in from one surface. Agents create inboxes, send, reply to, forward, and claim messages, draft and schedule mail, and wait for the next message to arrive, while threads, contacts, and search organise what comes in. Custom domains are added with DKIM rotation and DNS records, deliverability bounce and complaint rates are reported, and suppressed addresses are listed and removed. A public agent directory resolves and searches agent cards and verifies signatures, and signed webhooks stream events with redelivery and secret rotation.

## For AI agents

Create inboxes and send, reply to, draft, and search email for AI agents, add DKIM-signed custom domains, and resolve the agent directory. Reads and public lookups are open; writes use a bearer key.

## Scope

Does not handle SMS, voice calls, or marketing campaign management. Use for agent email inboxes, messaging, and directory lookup only.

## Capabilities

- Create inboxes and send, reply to, and forward email messages for AI agents
- Draft, schedule, and send messages, with idempotency keys on send, reply, and draft send
- Add custom domains, rotate DKIM keys, and read a domain's DNS records
- Resolve and search the agent directory and verify an agent's signature
- Register webhooks for events, redeliver past deliveries, and rotate signing secrets
- Read deliverability bounce and complaint rates and remove suppressed addresses

## Use cases

### Agent Email Handling via Jentic

An AI agent that corresponds by email can run its inbox through Agentboxd rather than a human mail client. Through Jentic the agent waits for the next message in an inbox, reads it, and sends a reply, using an idempotency key so a retried send does not go out twice, while the directory lets it verify who it is talking to.

Example prompt: Wait for the next message in an inbox, read its contents, and send a reply with an idempotency key so a retry does not resend it

### Custom-Domain Agent Mail

Teams send agent mail from their own domain with proper authentication. Agentboxd adds a custom domain, exposes its DNS records and a zone file, rotates and activates DKIM keys, and reports deliverability bounce and complaint rates, so an agent can set up a sending domain and monitor its health over time.

Example prompt: Add a custom domain, read its DNS records, start a DKIM rotation, and check deliverability bounce and complaint rates once it is verified

### Agent Directory and Discovery

Agents find and trust each other through a public directory. Agentboxd resolves an address or handle to an agent card, searches the directory, verifies an agent's signature, and exposes public agent cards, keys, and A2A records, so an agent can discover a counterpart and confirm its identity before messaging it.

Example prompt: Search the agent directory for a handle, resolve it to an agent card, and verify the agent's signature before sending a message

## Key endpoints

| Method | Path | Description |
| --- | --- | --- |
| POST | `/v1/inboxes` | Create an inbox |
| GET | `/v1/inboxes` | List inboxes |
| POST | `/v1/inboxes/{id}/messages/send` | Send a message |
| POST | `/v1/inboxes/{id}/messages/{messageId}/reply` | Reply to a message |
| GET | `/v1/messages` | List messages across inboxes |
| POST | `/v1/domains` | Add a custom domain |
| POST | `/v1/webhooks` | Create a webhook |

## Key resources

- **Inboxes and messages** — Create inboxes and send, reply to, forward, claim, draft, schedule, and search messages.
- **Domains** — Add custom domains, rotate and activate DKIM keys, verify DNS, and read zone files.
- **Directory** — Resolve and search agent cards, verify signatures, and read public keys and A2A records.
- **Webhooks and deliverability** — Register webhooks, redeliver events, rotate secrets, and read bounce and complaint rates.

## AI readiness

This API is usable in Jentic One now. Its AI-readiness score against Jentic's framework shows where it stands today and where improvements would make it even easier for agents to use.

- **Score:** 61 / 100
- **Maturity:** AI-Aware
- **Dimensions:**
  - Foundational Compliance: 74 / 100
  - Developer Experience & Jentic Compatibility: 69 / 100
  - AI-Readiness & Agent Experience: 46 / 100
  - Agent Usability: 72 / 100
  - Security: 60 / 100
  - AI Discoverability: 58 / 100
- **View full report:** https://jentic.com/apis/agentboxd.com/agentboxd/scorecard
- **How the score is calculated:** https://docs.jentic.com/reference/api-readiness-framework/overview/
- **More about the dimensions:** https://docs.jentic.com/reference/api-readiness-framework/specification/#dimensional-model-overview

### Score it yourself

Every API in the directory is allowlisted, so you can re-score it with no key required.

- **Score your own API:** https://jentic.com/scorecard.md
- **Scoring CLI agent skill:** https://github.com/jentic/jentic-api-scorecard/blob/main/skills/jentic-api-scorecard/SKILL.md

```sh
npx @jentic/api-scorecard-cli score <openapi-url>
```

## Why Jentic

- **Setup:** Wiring Agentboxd by hand means learning its bearer key, verifying custom domains and rotating DKIM keys, and sending an Idempotency-Key on send, reply, and draft send so a retry does not double-send. Through Jentic you install once, import Agentboxd from the API Directory, store the key once, and your agent sends and reads mail.
- **Permission scoping:** Agentboxd puts the inbox id in the URL path (`/v1/inboxes/{id}`), so a rule can pin your agent to one inbox and the operations you choose, such as sending and replying. You decide which operations it may call, so deleting inboxes or firing the workspace emergency stop stays out unless you add it.
- **Credential handling:** Your Agentboxd key is stored once, encrypted, by your own Jentic One instance and injected at execution time. It never enters the agent's prompt, logs, or context.
- **Discovery method:** Agents search Jentic by intent such as 'send an email from an agent inbox' or 'reply to a message', and Jentic returns the matching Agentboxd operation with its input schema so the agent calls the right endpoint directly.

## Related APIs

- **SendGrid Mail** — SendGrid sends transactional and marketing email at scale; Agentboxd adds agent inboxes, two-way messaging, and a directory.
- **Resend** — Resend is a developer-focused email sending API; Agentboxd layers inboxes, threads, and agent identity on top of sending.
- **Mailchimp** — Mailchimp runs marketing campaigns to lists; Agentboxd handles the transactional and agent-to-agent mail alongside them.

## FAQ

### What authentication does the Agentboxd API use?

The Agentboxd API authenticates with a key sent as a bearer token in the Authorization header, per its OpenAPI spec. Public directory endpoints and the signup challenge work without a key, while inbox, message, and domain operations carry it. Through Jentic the key is stored encrypted on your own instance and injected only on the operations that need it.

### Is there an Agentboxd MCP server?

You don't need an MCP server to give your agent Agentboxd. Jentic connects it directly from the API Directory: import Agentboxd, store your key once, and your agent sends, replies, and searches mail straight away. Operations are discovered on demand, so no extra server's tool definitions sit in your agent's context.

### Can I limit what my agent is allowed to do with the Agentboxd API?

Yes. Because the inbox id travels in the URL path, you can write a rule that allows only sending and replying on one inbox, so the agent handles that inbox and nothing else. Deleting inboxes or firing the workspace emergency stop stays out of the rule unless you add it, and every call the agent makes is logged.

### Can I send and reply to email with the Agentboxd API?

Yes. An agent sends a message from an inbox, replies to or forwards an existing message, and can draft and schedule mail for later. Send, reply, and draft send honour an Idempotency-Key header, and an agent can wait for the next message to arrive rather than polling the inbox.

### What are the rate limits for the Agentboxd API?

The Agentboxd OpenAPI spec does not publish numeric request rate limits. For incoming mail it recommends waiting for the next message or opening the realtime event stream instead of tight polling, which keeps request volume down. Check the Agentboxd documentation at https://agentboxd.com/docs/api for current limits.

### How do I send an email with the Agentboxd API through Jentic?

Search the Jentic API Directory for an intent like 'send an email from an agent inbox', import Agentboxd, and store your key once. Your agent then sends from an inbox and reads replies through Jentic. To run it on your own infrastructure, install Jentic One from its GitHub repo.