---
title: "Download one file from an agent's disk"
method: GET
path: "/v1/agents/{agentID}/files/download"
---

# Download one file from an agent's disk

`GET /v1/agents/{agentID}/files/download`

## Request

- Base URL: `https://api.plori.ai`
- URL: `https://api.plori.ai/v1/agents/{agentID}/files/download`
- Auth: HTTP bearer

Streams a file's bytes — text or binary, any size the account can hold — with its own content type for known extensions and `application/octet-stream` otherwise. This is the route a file link in an agent's reply resolves to. The disk is served directly, so an idle agent's files are downloadable with no running pod.

## Path parameters

- `agentID` string, uuid, required

## Query parameters

- `path` string, required

## Response `200`

The file's bytes. `Content-Disposition` is always an attachment, and `X-File-Size` carries the size the disk reported before the stream started.

## Other responses

- `400`: No path was given, or the path is a directory (`error: "is a directory"`) — list it instead.
- `401`: Missing or invalid credential.
- `403`: The resource belongs to another account.
- `404`: No such resource.

---

[API](https://skmtc.dev/plori/apis/plori-rest-api.md) · [All operations](https://skmtc.dev/plori/apis/plori-rest-api/llms.txt) · [OpenAPI document](https://skmtc.dev/plori/apis/plori-rest-api/revisions/15c4f9260dff?raw)