How to use free tier requests with Microsoft Foundry?

Priyank Panchal 20 Reputation points
2025-12-03T00:18:26.5633333+00:00

I created Azure Speech service for my speech-to-text transcription with F0 SKU and used fast transcription API endpoints which offers 5 hrs of free audio transcription/month. I was never charged for using this. I thought of using Microsoft Foundry instead since I plan to use translation in the future too. But, after a few trials, I see accumulated cost in my subscription against these services even though the requests are well under the free quota limits.

My foundry is in S0 tier, but I cannot find where to change this as it did not give SKU option while creating the foundry.

How can I use free tier with Microsoft Foundry? Here are the pricing links I am referring to:

https://azure.microsoft.com/en-us/pricing/details/cognitive-services/#pricing

https://azure.microsoft.com/en-us/pricing/details/cognitive-services/speech-services/

https://azure.microsoft.com/en-us/pricing/details/cognitive-services/translator/

Azure AI Speech
Azure AI Speech
An Azure service that integrates speech processing into apps and services.
0 comments No comments
{count} votes

Answer accepted by question author
  1. Nikhil Jha (Accenture International Limited) 4,150 Reputation points Microsoft External Staff Moderator
    2025-12-03T10:29:50.3133333+00:00

    Hello Priyank Panchal,

    Good day.

    I understand you are looking for to use free tier with Microsoft Foundry. The pricing pages you referenced are correct, but they apply to direct service usage, not Foundry hub usage.

    To directly answer your query: Foundry does not support F0 SKU because it is designed for production-grade orchestration and multi-service integration.

    The reason:

    • Free tier (F0 SKU) applies to individual Cognitive Services resources like Speech or Translator when created directly in Azure.
    • Microsoft Foundry is a consolidated orchestration platform that uses standard (S0) pricing for integrated services.
    • Foundry provisions services under S0 tier by default, which is billable from the first request.
    • Even if underlying services (Speech, Translator) have free quotas when used standalone, Foundry does not inherit those free-tier benefits.
    • This is documented in Foundry’s pricing model: orchestration and multi-agent capabilities are premium features.

    Workaround:

    If you want to leverage free tier benefits, you must:

    • Create individual Cognitive Services resources (e.g., Speech with F0 SKU) outside Foundry.
    • Call those resources directly via their REST APIs or SDKs instead of routing through Foundry.

    Suggestion if continued using SO tier in foundry:


    I hope this gives you a way ahead. Kindly Accept and Upvote the answer for other community members looking for similar remediation.

    1 person found this answer helpful.

1 additional answer

Sort by: Most helpful
  1. Q&A Assist
    2025-12-03T00:18:37.99+00:00
    AI generated content. This question contains an answer created with AI from Q&A Assist. Learn more

    To use the free tier with Microsoft Foundry, you need to ensure that you are using the correct pricing tier for your Speech service. The free tier (F0) is available for the Speech service, but it seems that your Foundry is currently set to the Standard (S0) tier, which incurs costs.

    Unfortunately, the free tier limits for the Speech service are not adjustable, and if you have created a Foundry resource under the S0 tier, it will not allow you to revert to the F0 tier directly. You may need to create a new Speech service resource specifically with the F0 SKU to utilize the free tier benefits.

    Additionally, the free tier for the Speech service allows for limited transactions and features, so ensure that your usage stays within those limits to avoid charges. For more detailed information on the quotas and limits for the Speech service, you can refer to the relevant documentation.


    References:

    0 comments No comments

Your answer

Answers can be marked as 'Accepted' by the question author and 'Recommended' by moderators, which helps users know the answer solved the author's problem.