> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.anthropod.in/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.anthropod.in/_mcp/server.

# Export data

> Create a CSV download using the existing dashboard export format.

Exports produce **CSV files** using the same column selection and flattening as dashboard downloads. Create a job, poll its status, and use `download_url` when it is completed.

#### Create

Send `POST /exports` with `resource` and `filters`. Receive 202 and the export ID.

#### Check progress

Call `GET /exports/{export_id}` until status is completed or failed.

#### Download CSV

Open the returned `download_url` before `download_expires_at`. Retrieve the job again for a fresh link while the file is available.

## Input

| Field                                    | Required      | Contract                                                                                                                                                                                           |
| ---------------------------------------- | ------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `resource`                               | Yes           | `conversations`, `customers`, or `voice_conversations`.                                                                                                                                            |
| `filters`                                | Yes           | Same resource, date and dashboard-filter meaning as list retrieval.                                                                                                                                |
| `filters.time_field`                     | No            | `conversation_time` for conversations and Voice Agents; `last_interaction_at` for customers. Defaults to the resource's field.                                                                     |
| `filters.start_time`, `filters.end_time` | Yes           | Integer epoch seconds; start inclusive, end exclusive; at most 31 days.                                                                                                                            |
| `filters.service_id`                     | For analytics | Required for conversations and customers. Voice Agent exports use optional `agent_id`.                                                                                                             |
| `filters.filter`                         | No            | Same optional dashboard filter object documented on the corresponding list API.                                                                                                                    |
| `format`                                 | No            | `csv` only; defaults to `csv`.                                                                                                                                                                     |
| `parameters`                             | No            | Dashboard download sections to include. Defaults to all supported sections; must include `Meta Data` for analytics or `Metadata` for Voice Agents. Allowed labels are listed in the API reference. |

## CSV contents

| Resource                  | Existing download sections                                                                                                                                                   |
| ------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Conversations             | Metadata, summary, Agent Feedback, Business Insights, dispositions, concerns/queries, Lead Quality, templates, Custom Module, explanations, subparameters and transcription. |
| Customers                 | Metadata, summary, Business Insights, dispositions, concerns/queries, Lead Quality, templates, Custom Module, timeline, explanations and subparameters.                      |
| Voice Agent conversations | Metadata, transcription, scheduling, CRM payload, request data, tool calls, subparameters, summary, Business Insights and explanations.                                      |

The CSV uses the dashboard's existing column names and handling of nested values. Columns depend on the chosen sections and configured analytics. Transcription text is included when `Transcript` is selected (the default includes it). Recording audio is not embedded in a CSV file.

## Output

The API returns a JSON job object in `data`; it does not embed the CSV content.

| Field                        | Format                                                                                                                   |
| ---------------------------- | ------------------------------------------------------------------------------------------------------------------------ |
| `id`, `resource`, `format`   | Export ID, selected resource, and `csv`.                                                                                 |
| `status`                     | `queued`, `processing`, `completed`, `failed`, or `expired`.                                                             |
| `created_at`, `completed_at` | Epoch seconds; `completed_at` is null while queued or processing and is set when processing finishes, including failure. |
| `download_url`               | Temporary signed CSV link when completed; otherwise null.                                                                |
| `download_expires_at`        | Link expiry in epoch seconds, or null.                                                                                   |
| `error`                      | Error code/message when failed; otherwise null.                                                                          |

## Limits

| Control           | Limit                                                                                       |
| ----------------- | ------------------------------------------------------------------------------------------- |
| Date window       | 31 days per job. Older stored data can be requested.                                        |
| File size         | Up to 100,000 records and 100 MB. Narrow the filters or date range if the job is too large. |
| Active jobs       | Five per key.                                                                               |
| File availability | Seven days after completion.                                                                |
| Link lifetime     | At most ten minutes; retrieve the job for a fresh link.                                     |

Use adjacent windows for longer periods: the previous end\_time becomes the next start\_time. Only the key that created the job can retrieve it. Expired, deleted or rotated keys cannot start downloads or obtain fresh links. Retain the returned job ID and poll it; submitting another request creates another job.