Create and isolate many inboxes, one per agent, run or customer
One company's docs show a full fit.
What counts as done
Inbox creation by API returns an address in one call, a retried create returns the same inbox, the inbox works at once, and a key scoped to one inbox cannot read another.
How agents ask for it: create an inbox for each of my agents
, give every customer's agent a separate address
, spin up a fresh inbox per test run
Does the job 1
The docs show it meets the test above, and nothing we read says otherwise.
Does the job, with a limit 7
It does the job, but something in the docs or a user's account will get in an agent's way. The limit is named in each row.
| Listing | Why |
|---|---|
| AgentMail API + MCP BB AgentMail, Inc. | client_id makes inbox creation idempotent and keys can be scoped to one pod or inbox. Open issue 14 (a20) shows a send 404ing for 1 to 2 seconds after create, and Free caps at 3 inboxes. source |
| Inbound C EXON ENTERPRISE LLC | POST /mailboxes creates a managed mailbox whose password is scoped to its addresses, with read or read_write access, but accounts default to 100 managed credentials, and create takes no client id for safe retries. source |
| Mailows C DW Technology LLC | POST /v1/inboxes creates a mailbox and a tenant-wide provisioning token issues each mailbox its own scoped token once, but Free has one agent mailbox, extra ones are $12 a year each, and create takes no idempotency key. source |
| mails.ai Agent Email D Mails.ai | Agents are created on first send and a key can be bound to one agent with chosen scopes, but Free allows one agent and Pro five; unlimited agents start at Scale ($99). source |
| MailSlurp B Pettman OÜ | CreateInboxWithOptions makes an inbox per test run with an expiry, and agent keys are checked against selected inbox IDs or tags, but Free allows 100 creations and 50 retained inboxes, and create takes no idempotency key in the docs read. source |
| Postfleet E Postfleet | Mailbox-bound keys can't reach other mailboxes and can_send false gives a reading key, but each mailbox create makes a new one (not idempotent) and plans cap mailboxes at 3, 5, 10 and 150. source |
| Robotomail B Tiny Bot Labs Limited | Mailbox-scoped keys restrict an agent to its own mailboxes, but Free and Solo have one mailbox, and the docs read show no idempotent create. source |
Doesn't do this job 4
Listings in the category that do something else, with the reason.
| Listing | Why |
|---|---|
| Cherami C Abderrahmane Gourragui (sole operator, Quebec) | An account gets 2 inboxes, more only by request, inbox creation takes a chosen local_part with no client id, and every key and connection reaches every inbox on the account. source |
| Clawpost E Frank Harris | The CLI can create exact-address inboxes, but one shared API key grants every operation on every inbox, and creation isn't idempotent; it isn't built for fleets. source |
| Hostinger Agentic Mail C Hostinger International Limited | The Mail API has no endpoint to create a mailbox; mailboxes are bought on Hostinger Email plans in hPanel. Tokens can be scoped to selected mailboxes. source |
| InboxAPI D Sitka Capital Pty Ltd | One account is one agent address created through the CLI; there is no API to create many inboxes or scope keys per inbox. source |
What stops agents finishing
- account-wide keys only, so every worker can read every inbox
- a cap on inboxes or managed credentials below the fleet size
- manual setup per address
Around it
Before: agent-auth.provision-identity. After: agent-inboxes.verification-code, agent-inboxes.own-identity.
Where this comes from
Weight 13 of 100 in Agent inbox APIs: 5 of 31 accounts: a platform with one address per agent (a09), a team of agents set up by hand (a07), DIY first then switched (a01), the fresh-inbox 404 (a20) and the fleet recommendation in a29
The same as data: /api/v1/use-cases/agent-inboxes.fleet-provisioning.json, or the MCP tool get_use_case with id agent-inboxes.fleet-provisioning. Other jobs in this category: /api/v1/niches?category=agent-inboxes.
For companies
Do agents find, use and choose your tools?
An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.