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

# AI Number API

Jentic publishes the only available OpenAPI specification for AI Number API, keeping it validated and agent-ready. AI Number, formerly Sendbee, is a WhatsApp Business messaging platform for managing teams, contacts, and conversations across customer support and outbound campaigns. The API exposes 20 endpoints covering team rosters, contact records with tags and custom fields, conversation threads, customer-care messages, and approved template messages. Authentication uses paired authToken and X-API-Key headers.

## For AI agents

Send WhatsApp messages, manage contacts and tags, and orchestrate customer-care conversations through the AI Number platform.

## Scope

Does not handle SMS, voice calls, or email - use for WhatsApp Business messaging and contact management only.

## Capabilities

- Send approved WhatsApp template messages to contacts via /conversations/messages/templates/send
- Send free-form customer-care replies inside the 24-hour service window
- Create, update, and tag contacts to drive segmentation and routing rules
- Define custom contact fields to capture business-specific attributes
- List and reassign conversation folders to balance team workload
- Pull message history for a conversation to feed AI summarisation

## Use cases

### WhatsApp Customer Support Automation

Operate a WhatsApp-first support inbox where agents and bots handle inbound conversations through the same API. List open threads with /conversations, pull full transcripts via /conversations/messages, then reply with /conversations/messages/send while the customer-care window is open. Outside the window, fall back to approved templates.

Example prompt: List the 5 oldest unread conversations, summarise each with their last 10 messages, and post a customer-care reply to the first one acknowledging receipt

### Outbound WhatsApp Campaigns with Templates

Send approved WhatsApp template messages to a tagged contact list - for order confirmations, appointment reminders, or re-engagement. The template send endpoint accepts variable substitutions per recipient, and tag-based contact filtering keeps the audience tight.

Example prompt: Find all contacts tagged 'reminder_due', then send the appointment_reminder template to each with their first name as the variable

### Contact Enrichment and Segmentation

Build a tagged, custom-field-rich contact database that doubles as a CRM lite for messaging. Create custom fields with /contacts/fields, tag contacts based on lifecycle stage, and update records as conversations evolve so subsequent messages can be targeted by attribute.

Example prompt: Create a custom field called 'plan_tier' of type string, then update contact c_777 setting plan_tier=premium and add the tag 'vip'

### AI Agent WhatsApp Operator

An agent triages an inbound WhatsApp conversation through Jentic - fetching the thread, drafting a reply, and posting it back to the customer without operator handoff. Jentic handles the auth pairing (authToken plus X-API-Key) so the agent only sees scoped operations.

Example prompt: Search Jentic for 'send whatsapp customer care message', load the schema for /conversations/messages/send, and post a reply to conversation conv_123 saying 'Your order has shipped'

## Key endpoints

| Method | Path | Description |
| --- | --- | --- |
| POST | /conversations/messages/send | Send a customer-care WhatsApp message |
| POST | /conversations/messages/templates/send | Send an approved WhatsApp template message |
| GET | /conversations | List conversation threads |
| GET | /conversations/messages | List messages within a conversation |
| POST | /contacts/subscribe | Create a new contact |
| GET | /contacts | List contacts |
| POST | /contacts/tags | Create a contact tag |

## Key resources

- **Teams** — Team and member rosters for routing conversations
- **Contacts** — Customer contact records with tags and custom fields
- **Contact Tags** — Labels used for contact segmentation and filtering
- **Contact Fields** — Custom attributes attached to contacts
- **Conversations** — WhatsApp conversation threads and folder state
- **Messages** — Customer-care messages, template messages, and message history

## Why Jentic

- **Setup:** Wiring AI Number by hand means sending two headers, an authToken and an X-API-Key, on every call to the Sendbee host, plus building message payloads and your own retry logic. Through Jentic you install once, import AI Number from the API Directory, store both values once, and your agent calls it.
- **Permission scoping:** AI Number carries its conversation and contact targets in the request body, so scope your agent by operation: limit it to the operations it needs, such as sending a template message, and leave out contact tagging or subscription endpoints it does not use.
- **Credential handling:** Your AI Number authToken and X-API-Key are stored once, encrypted, by your own Jentic One instance and injected at execution time. They never enter the agent's prompt, logs, or context.
- **Discovery method:** Agents search Jentic by intent such as 'send a whatsapp template', and Jentic returns the matching AI Number operation with its input schema so the agent calls the right endpoint without browsing the Sendbee reference docs.

## Related APIs

- **Gupshup API** — Multi-channel conversational messaging including WhatsApp Business, RCS, and SMS
- **Twilio Messaging API** — WhatsApp, SMS, and MMS messaging with global reach and detailed delivery telemetry
- **HubSpot CRM Contacts** — Sync AI Number contacts with a primary CRM for unified customer records

## FAQ

### Why is there no official OpenAPI spec for AI Number API?

AI Number does not publish an OpenAPI specification. Jentic generates and maintains this spec so that AI agents and developers can call AI Number API via structured tooling. It is validated against the live API and kept up to date. Get started with Jentic One, the self-hosted execution layer.

### What authentication does the AI Number API use?

AI Number requires two paired headers: authToken and X-API-Key. Both are stored encrypted in your Jentic One instance, so agents calling AI Number through Jentic receive scoped access without ever seeing the raw credentials.

### Can I send a WhatsApp template message with the AI Number API?

Yes. POST /conversations/messages/templates/send dispatches an approved template to a contact, with variable substitutions for personalised content. Use GET /conversations/messages/templates first to confirm which templates are approved.

### What are the rate limits for the AI Number API?

Rate limits are not declared in the spec and are governed by the underlying WhatsApp Business platform. For high-throughput sending, batch by contact tag and respect WhatsApp's per-template messaging quality rating.

### How do I send a customer-care reply with the AI Number API through Jentic?

Run pip install jentic, search Jentic for 'send whatsapp customer care message', load the schema for POST /conversations/messages/send, then execute with the conversation ID and message body. Jentic injects both the authToken and X-API-Key headers.

### Can I tag and segment contacts through the AI Number API?

Yes. Create tags with POST /contacts/tags and custom fields with POST /contacts/fields, then PUT /contacts to update individual records. Tags and fields can both be used as filters when targeting outbound campaigns.

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

Yes. Because you run Jentic One yourself, your own rules decide which AI Number operations and credentials the agent can use, and the agent only ever sees the operations you grant. AI Number carries its conversation and contact targets in the request body, so you scope the agent by operation rather than by object. For example, you can allow only POST /conversations/messages/templates/send so the agent sends approved templates, while leaving out contact tagging (POST /contacts/tags) and contact creation (POST /contacts/subscribe) endpoints it does not need.