How to use Audio upload
Uploads a recording and selects language, channel mapping, diarization and processing profile.
What this section is for
Uploads a recording and selects language, channel mapping, diarization and processing profile.
Before you start
You need access to IAM.Speech, the correct tenant/workspace and permission to change this configuration. Prepare a non-sensitive example value such as call-demo.wav.
Buttons and forms
| Control | What it does |
|---|---|
| Choose file | Performs the action named in the interface. Review the resulting state or message before continuing. |
| Language | Performs the action named in the interface. Review the resulting state or message before continuing. |
| Diarization | Performs the action named in the interface. Review the resulting state or message before continuing. |
| Upload | Performs the action named in the interface. Review the resulting state or message before continuing. |
| Cancel | Stops or pauses the current operation without deleting its configuration. |
Primary path in the interface
- Open the application link above and sign in with the least-privileged role required for this section.
- Select the intended tenant or workspace. Do not continue while the header shows a different tenant.
- In the navigation, click Audio upload. The expected address is https://app.iamspeech.ru/upload.
- Click Upload. Enter the documented safe values:
call-demo.wav. Do not paste a production secret into a free-text field. - Review every field before the final Save, Run, Publish or Confirm action. If a preview or validation control exists, run it first.
- Review the validation message or changed row. The result must match this purpose: Uploads a recording and selects language, channel mapping, diarization and processing profile.
- Only then continue to the next dependent section. Copy the request/correlation ID when the interface shows one.
Second path: direct link and CLI/API
- Open the deep link https://app.iamspeech.ru/upload instead of navigating through the sidebar.
- Confirm the tenant/workspace after the page loads; a deep link must not silently switch scope.
- Before automating the API path, run this product-specific read-only or validation command. It proves reachability but does not bypass application permissions:
export IAM_SPEECH_URL=https://api.iamspeech.ru
curl --fail --silent --show-error "$IAM_SPEECH_URL/healthz"
How to verify the result
The pipeline reaches a final state, duration matches the file and transcript is available.
If it did not work
| Symptom | What to do |
|---|---|
| 401 / sign-in screen | Sign in again, then return to this page. Do not put a session token in a URL. |
| 403 | Check the selected tenant and request the documented role; retrying with the same role will not help. |
| 404 | Verify the exact path and whether the feature is enabled in this environment. |
| 409 | Search for an object named call-demo.wav; continue with it or choose a unique name. |
| 422 / validation error | Correct only the highlighted fields. Keep identifiers lowercase unless the form explicitly allows otherwise. |
| 429 | Wait for the displayed retry interval or reduce the request scope; do not start parallel retries. |
| 5xx / timeout | Refresh once after the server has recovered; do not repeat a destructive action while its result is unknown. Save the correlation ID and timestamp for support. |