nprasad2077/mcp-microsoft
1
0 Fork 0
mirror of https://github.com/microsoft/mcp.git synced 2026-11-04 12:01:31 +00:00
Code Issues Packages Projects Releases Wiki Activity

docs: clarify Tool Description Evaluator embeddings endpoint setup (#4024)

Browse Source
This commit is contained in:
KarishmaGhiya authored and GitHub committed 2026-10-23 00:45:21 +00:00
1 parent f1a5b8dfe1
commit 130fc0f2f0
2 files changed
+21 -3
Download Patch File Download Diff File Expand all files Collapse all files

No files matched your search

eng/tools/ToolDescriptionEvaluator/Quickstart.md
+9 -1
View File
@@ -22,7 +22,15 @@ An Azure OpenAI deployment of the text embedding model `text-embedding-3-large`.
### Minimal Setup
Set your Azure OpenAI endpoint and API key as environment variables:
To find the complete embeddings URL for your Azure OpenAI resource in Microsoft Foundry:
1. Open your resource and select **Deployments** under **Shared resources**.
2. Select your `text-embedding-3-large` deployment, not a chat model deployment.
3. On the deployment details page, find the **Endpoint** section and copy **Target URI**. This is the complete URL, including the deployment path, `/embeddings`, and the `api-version` query parameter.
Set `AOAI_ENDPOINT` to this **Target URI**, not the resource's base endpoint (such as `https://<your-resource>.openai.azure.com/`). The evaluator sends requests directly to the configured URL.
Set the complete embeddings URL and an API key for the same resource as environment variables:
```bash
export AOAI_ENDPOINT="https://<your-resource>.openai.azure.com/openai/deployments/<embeddings-deployment-name>/embeddings?api-version=<api-version>"
eng/tools/ToolDescriptionEvaluator/README.md
+12 -2
View File
@@ -189,8 +189,18 @@ This application requires two environment variables to be configured:
#### Required Environment Variables
1. **`AOAI_ENDPOINT`** - Your Azure OpenAI endpoint base URL (without deployment path)
2. **`TEXT_EMBEDDING_API_KEY`** - Your Azure OpenAI API key
1. **`AOAI_ENDPOINT`** - The complete Azure OpenAI embeddings URL, including the deployment path, `/embeddings`, and the `api-version` query parameter
2. **`TEXT_EMBEDDING_API_KEY`** - An Azure OpenAI API key for the same resource
#### Find the Embeddings URL in Microsoft Foundry
To find the complete embeddings URL for your Azure OpenAI resource in Microsoft Foundry:
1. Open your resource and select **Deployments** under **Shared resources**.
2. Select your `text-embedding-3-large` deployment, not a chat model deployment.
3. On the deployment details page, find the **Endpoint** section and copy **Target URI**.
Use this **Target URI** as `AOAI_ENDPOINT`. Do not use only the resource's base endpoint (such as `https://<your-resource>.openai.azure.com/`): the evaluator sends requests directly to the configured URL without appending the deployment path or operation.
#### Option 1: Environment Variables (Recommended for Production)