***

title: Resource model
subtitle: The live contact resource shape and the conventions it follows
slug: resource-model
--------------------

## Live resources today

| Resource            | Live operations | Identifier prefix             |
| ------------------- | --------------- | ----------------------------- |
| Contact             | List, get       | `cnt_`                        |
| Company             | List, get       | `cmp_`                        |
| User                | List, get       | `usr_`                        |
| Tag                 | List, get       | `tag_`                        |
| Engagement          | List, get       | `ngmt_`                       |
| Transcript timeline | Get             | n/a (nested under engagement) |

## Shared conventions

* Responses use `snake_case`
* Timestamps use ISO 8601 UTC strings
* Most resources include `created_at` and `updated_at`
* The contact payload can include lightweight related-resource references such as `{ "id": "cmp_...", "object": "company" }`
* All list resources use the same cursor-paginated list envelope and default to `updated_at` descending order
* Transcript timeline entries expose `speaker`, `text`, `start_seconds`, and `end_seconds`

## Contact fields

* Identity: `id`, `object`
* Names: `first_name`, `last_name`
* Profile: `description`, `time_zone`
* Relationships: `company`
* Communication channels: `emails`, `phone_numbers`
* Audit timestamps: `created_at`, `updated_at`

## Resource highlights

* Company fields include `domains`, `industry`, `website`, and `number_of_employees`
* User fields include safe profile metadata such as `job_title`, `emails`, and `last_activity_at`
* Tag fields include `hex_color` and `auto_tagging_enabled`
* Engagement fields include `engagement_type`, `data_source`, `processing_status`, and timing metadata
* Transcript timeline speakers can resolve to `user`, `contact`, or `unknown`