Hey Amal, working with the Adobe Document Status API I was able to handle their response like this!

Via Postman, the response was multipart with the body of the response in 1 JSON and 1 octet-stream.
(header Content-Type: multipart/mixed; boundary=Boundary_583415_1634289954_1663190050427;charset=UTF-8)

--Boundary_583415_1634289954_1663190050427
Content-Type: application/json
Content-Disposition: form-data; name="contentAnalyzerResponse"

{"cpf:inputs": … :{"documentOut":{"cpf:location":"multipartLabel","dc:format":"application/pdf"}}}
--Boundary_583415_1634289954_1663190050427
Content-Type: application/octet-stream
Content-Disposition: form-data; name="multipartLabel"

%PDF-1.7
…
%%EOF

--Boundary_583415_1634289954_1663190050427--

HTTPoison ships a function to parse multipart bodies HTTPoison.Handlers.Multipart.decode_body(response). From there we can parse the file body and then write that to disk to read it!
Hex Doc: decode_body/1

with {:ok, %{status_code: code} = response} when code in [200, 202] <- HTTPoison.get(url, headers),
     {:decode, parts} when is_list(parts) <- {:decode, HTTPoison.Handlers.Multipart.decode_body(response)},
     {_headers, file_body} <- Enum.find(parts, fn {headers, _body} -> headers |> Map.new() |> Map.get("Content-Type") == "application/octet-stream" end) do
  File.write!("./test.pdf", file_body)
end