> ## Documentation Index
> Fetch the complete documentation index at: https://support.configview.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Plaud setup

Configview reads your Plaud Team workspace (the AI voice recorder) through the **Plaud Developer Platform enterprise API** (`https://api.plaud.ai/enterprise/v1`), using a **client id and an RSA private key** created for your workspace. Configview only sends `GET` requests. It never reads transcripts, summaries or audio, and it never deletes, edits or writes back anything in Plaud.

You will end up with **3 required secrets** (`PLAUD_CLIENT_ID`, `PLAUD_PRIVATE_KEY`, `PLAUD_JWT_AUDIENCE`) and 2 optional ones in Configview when setup is complete.

> **Private beta.** The Plaud Developer Platform enterprise API is a **private beta**. Access is requested through the contact form at [https://dev.plaud.ai](https://dev.plaud.ai), reviewed by Plaud, and only then do you get a console to create the credential. It covers one Team workspace; Personal workspaces are never readable. The API may change before general availability.

> **Scope of this integration today.** Configview answers "who is in our Plaud workspace and with what role, who has left but still owns recordings, how much is being recorded and by whom, and who deleted what": the workspace, every member with role (admin, member, guest) and active flag, recording **metadata** for the last 90 days (owner, recorded date, duration, whether a transcript and a summary exist, status) and deletion records. File names, transcripts, summaries and audio are **never** requested or stored.

***

## What you need

* A Plaud **Team workspace** and a workspace admin
* Access to the Plaud Developer Platform (private beta, see above)
* A way to generate an RSA key pair (for example `openssl genrsa -out plaud-key.pem 2048` and `openssl rsa -in plaud-key.pem -pubout -out plaud-key.pub`)

***

## Step 1: Open the Plaud page in Configview

Open Configview in a second browser tab and leave it open:

`https://{companyname}.configview.com/admin/integrations/plaud`

***

## Step 2: Create the credential in Plaud

1. Request access at [https://dev.plaud.ai](https://dev.plaud.ai) and wait for approval
2. In the Plaud developer console create a client for your Team workspace and register the **public** key (RSA, 2048 bits or more). Choose an expiry of 180 days or less
3. Grant **only read scopes**: `members:read` and `files:read`. The deletion list below needs `files:delete`, see "Deletion records"
4. Note the **client id** (`client_...`), the **key id** (`kid`, if shown) and the **audience** (the issuer string Plaud gives you with the credential)
5. In the Configview tab save, under **Credentials**: the client id as `PLAUD_CLIENT_ID`, the whole private key file (including the `BEGIN` and `END` lines) as `PLAUD_PRIVATE_KEY`, and the audience as `PLAUD_JWT_AUDIENCE`. Optional: `PLAUD_KEY_ID` (leave empty with a single active key) and `PLAUD_API_BASE` (only for Japan workspaces)

At the start of every run Configview signs a short client assertion with your private key (RS256) and exchanges it at `https://api.plaud.ai/team-oauth/oauth/token` for a one-hour access token. Nothing rotates and no refresh token exists. Revoking the key in Plaud stops Configview; replace the key before its expiry.

***

## Step 3: Connect and verify

1. Click **Connect**. Configview creates its tables and schedules every collector at your default run time:

| Script | Notes |
| - | - |
| **Workspaces** | The workspace bound to the credential. |
| **Members** | Everyone in the workspace, including people who left, with role and active flag. |
| **Files** | Recording metadata for the last 90 days. |
| **Deletion records** | Who deleted which recording (last 90 days). Skipped when the credential lacks `files:delete`. |

2. Click **Verify now**. The health check reads the secrets, checks that the private key signs, exchanges an assertion for a token and lists one workspace.

If a check fails:

* **Token endpoint 400 or 401.** The client id, audience or key id is wrong, the key is not the one registered, or the key expired.
* **403.** The credential lacks `members:read` or `files:read`, or the beta access was not granted for this workspace.

***

## Data Tables

All four tables keep only the newest run (snapshots).

| Table | Source | Key Columns |
| - | - | - |
| `plaud_workspaces` | `GET /team/workspaces` | workspace\_id, name, active, created\_at |
| `plaud_members` | `GET /team/members` | member\_id, email, role, active, joined\_at |
| `plaud_files` | `GET /team-files/items` | file\_id, created\_by\_member\_id, status, recorded\_at, duration\_ms, has\_transcript, has\_summary |
| `plaud_audit_logs` | `GET /team-files/deletion-records` | record\_id, file\_id, owner, method, actor\_client\_id, status, requested\_at, completed\_at |

***

## Things worth knowing

**Whose recordings are visible.** The file feed returns Team Files plus the private files of members who **consented to API access**. A member who has not consented contributes no files, so recording counts can be lower than reality.

**Metadata only.** The feed returns the file name; Configview does not read it and does not store it. Transcript, summary and audio endpoints are never called.

**Members who left.** Configview lists members with `status=all`, so people who left or were removed appear with `active = 0` and their recordings still resolve to a name. Invited and suspended members are also `active = 0`. Plaud exposes the join date but no last-active date.

**Files are re-read in full.** The feed is a change feed; Configview does not keep a checkpoint between runs, so each run reads the whole feed, keeps the newest version of each file and stores the last 90 days (`recorded_at`, else `created_at`).

**Deletion records.** Listing deletion records needs the `files:delete` scope, which also allows deleting files. Configview never calls a delete endpoint, but if you do not want to grant that scope, leave it off: the Deletion records collector then skips itself and the other three keep working. Per-artifact detail and retention text are not stored.

**Rate limit.** 120 requests per minute per client for the `/team` and `/team-files` endpoint groups each, 100 rows per page. Configview paces itself at 100 per minute and waits when Plaud answers 429.

**Regions.** US workspaces use `api.plaud.ai`. Japan workspaces receive their own host with their credential: set `PLAUD_API_BASE` to `https://<host>/enterprise/v1`.

## Known gaps

* No seat or plan counts, last-active date, 2FA or SSO flags, API-key listing, webhooks or integrations are exposed by the enterprise API
* Parent workspaces (`parent_id`) are always empty in v1
* Sharing or privacy flags per recording are not exposed

## What isn't collected

* File names, transcripts, summaries, notes and audio (and their download links)
* Deletion artifact detail and retention statements
* Private keys, client assertions and access tokens


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.